Added current engine clock and current memory clock to the status display (ADL only...
[hashcat.git] / src / hashcat.c
index de05dd3..1141815 100644 (file)
@@ -38,7 +38,7 @@ double TARGET_MS_PROFILE[4]     = { 2, 12, 96, 480 };
 #define RESTORE_DISABLE         0
 #define STATUS                  0
 #define STATUS_TIMER            10
-#define STATUS_AUTOMAT          0
+#define MACHINE_READABLE        0
 #define LOOPBACK                0
 #define WEAK_HASH_THRESHOLD     100
 #define SHOW                    0
@@ -76,7 +76,7 @@ double TARGET_MS_PROFILE[4]     = { 2, 12, 96, 480 };
 #define BITMAP_MAX              24
 #define GPU_TEMP_DISABLE        0
 #define GPU_TEMP_ABORT          90
-#define GPU_TEMP_RETAIN         80
+#define GPU_TEMP_RETAIN         70
 #define WORKLOAD_PROFILE        2
 #define KERNEL_ACCEL            0
 #define KERNEL_LOOPS            0
@@ -349,9 +349,7 @@ const char *USAGE_BIG[] =
   "",
   "Usage: %s [options]... hash|hashfile|hccapfile [dictionary|mask|directory]...",
   "",
-  "###########",
-  "# Options #",
-  "###########",
+  "- [ Options ] -",
   "",
   " Options Short / Long          | Type | Description                                          | Example",
   "===============================|======|======================================================|=======================",
@@ -366,7 +364,7 @@ const char *USAGE_BIG[] =
   "     --force                   |      | Ignore warnings                                      |",
   "     --status                  |      | Enable automatic update of the status-screen         |",
   "     --status-timer            | Num  | Sets seconds between status-screen update to X       | --status-timer=1",
-  "     --status-automat          |      | Display the status view in a machine readable format |",
+  "     --machine-readable        |      | Display the status view in a machine readable format |",
   "     --loopback                |      | Add new plains to induct directory                   |",
   "     --weak-hash-threshold     | Num  | Threshold X when to stop checking for weak hashes    | --weak=0",
   "     --markov-hcstat           | File | Specify hcstat file to use                           | --markov-hc=my.hcstat",
@@ -436,12 +434,10 @@ const char *USAGE_BIG[] =
   "     --increment-min           | Num  | Start mask incrementing at X                         | --increment-min=4",
   "     --increment-max           | Num  | Stop mask incrementing at X                          | --increment-max=8",
   "",
-  "################",
-  "## Hash modes ##",
-  "################",
+  "- [ Hash modes ] -",
   "",
   "      # | Name                                             | Category",
-  "  ------+--------------------------------------------------+--------------------------------------",
+  "  ======+==================================================+======================================",
   "    900 | MD4                                              | Raw Hash",
   "      0 | MD5                                              | Raw Hash",
   "   5100 | Half MD5                                         | Raw Hash",
@@ -654,12 +650,10 @@ const char *USAGE_BIG[] =
   "  12700 | Blockchain, My Wallet                            | Password Managers",
   "  13400 | Keepass 1 (AES/Twofish) and Keepass 2 (AES)      | Password Managers",
   "",
-  "#####################",
-  "## Outfile Formats ##",
-  "#####################",
+  "- [ Outfile Formats ] -",
   "",
   "  # | Format",
-  " ---+--------",
+  " ===+========",
   "  1 | hash[:salt]",
   "  2 | plain",
   "  3 | hash[:salt]:plain",
@@ -676,35 +670,29 @@ const char *USAGE_BIG[] =
   " 14 | plain:hex_plain:crack_pos",
   " 15 | hash[:salt]:plain:hex_plain:crack_pos",
   "",
-  "##########################",
-  "## Rule Debugging Modes ##",
-  "##########################",
+  "- [ Rule Debugging Modes ] -",
   "",
   "  # | Format",
-  " ---+--------",
+  " ===+========",
   "  1 | Finding-Rule",
   "  2 | Original-Word",
   "  3 | Original-Word:Finding-Rule",
   "  4 | Original-Word:Finding-Rule:Processed-Word",
   "",
-  "##################",
-  "## Attack Modes ##",
-  "##################",
+  "- [ Attack Modes ] -",
   "",
   "  # | Mode",
-  " ---+------",
+  " ===+======",
   "  0 | Straight",
   "  1 | Combination",
   "  3 | Brute-force",
   "  6 | Hybrid Wordlist + Mask",
   "  7 | Hybrid Mask + Wordlist",
   "",
-  "#######################",
-  "## Built-in Charsets ##",
-  "#######################",
+  "- [ Built-in Charsets ] -",
   "",
   "  ? | Charset",
-  " ---+---------",
+  " ===+=========",
   "  l | abcdefghijklmnopqrstuvwxyz",
   "  u | ABCDEFGHIJKLMNOPQRSTUVWXYZ",
   "  d | 0123456789",
@@ -712,22 +700,18 @@ const char *USAGE_BIG[] =
   "  a | ?l?u?d?s",
   "  b | 0x00 - 0xff",
   "",
-  "#########################",
-  "## OpenCL Device Types ##",
-  "#########################",
+  "- [ OpenCL Device Types ] -",
   "",
   "  # | Device Type",
-  " ---+-------------",
+  " ===+=============",
   "  1 | CPU",
   "  2 | GPU",
   "  3 | FPGA, DSP, Co-Processor",
   "",
-  "#######################",
-  "## Workload Profiles ##",
-  "#######################",
+  "- [ Workload Profiles ] -",
   "",
   "  # | Performance | Runtime | Power Consumption | Desktop Impact",
-  " ---+-------------+---------+-------------------+----------------",
+  " ===+=============+=========+===================+=================",
   "  1 | Low         |   2 ms  | Low               | Minimal",
   "  2 | Default     |  12 ms  | Economic          | Noticeable",
   "  3 | High        |  96 ms  | High              | Unresponsive",
@@ -772,7 +756,7 @@ static double get_avg_exec_time (hc_device_param_t *device_param, const int last
   return exec_ms_sum / exec_ms_cnt;
 }
 
-void status_display_automat ()
+void status_display_machine_readable ()
 {
   FILE *out = stdout;
 
@@ -930,9 +914,9 @@ void status_display ()
   if (data.devices_status == STATUS_STARTING) return;
   if (data.devices_status == STATUS_BYPASS)   return;
 
-  if (data.status_automat == 1)
+  if (data.machine_readable == 1)
   {
-    status_display_automat ();
+    status_display_machine_readable ();
 
     return;
   }
@@ -1565,21 +1549,25 @@ void status_display ()
       {
         char utilization[HM_STR_BUF_SIZE] = { 0 };
         char temperature[HM_STR_BUF_SIZE] = { 0 };
-        char fanspeed[HM_STR_BUF_SIZE] = { 0 };
+        char fanspeed[HM_STR_BUF_SIZE]    = { 0 };
+        char corespeed[HM_STR_BUF_SIZE]   = { 0 };
+        char memoryspeed[HM_STR_BUF_SIZE] = { 0 };
 
-        hm_device_val_to_str ((char *) utilization, HM_STR_BUF_SIZE, "%", hm_get_utilization_with_device_id (device_id));
-        hm_device_val_to_str ((char *) temperature, HM_STR_BUF_SIZE, "c", hm_get_temperature_with_device_id (device_id));
+        hm_device_val_to_str ((char *) utilization, HM_STR_BUF_SIZE,   "%", hm_get_utilization_with_device_id (device_id));
+        hm_device_val_to_str ((char *) temperature, HM_STR_BUF_SIZE,   "c", hm_get_temperature_with_device_id (device_id));
+        hm_device_val_to_str ((char *) corespeed,   HM_STR_BUF_SIZE, "Mhz", hm_get_corespeed_with_device_id   (device_id));
+        hm_device_val_to_str ((char *) memoryspeed, HM_STR_BUF_SIZE, "Mhz", hm_get_memoryspeed_with_device_id (device_id));
 
-        if (device_param->vendor_id == VENDOR_ID_AMD)
+        if (device_param->device_vendor_id == VENDOR_ID_AMD)
         {
           hm_device_val_to_str ((char *) fanspeed, HM_STR_BUF_SIZE, "%", hm_get_fanspeed_with_device_id (device_id));
         }
-        else if (device_param->vendor_id == VENDOR_ID_NV)
+        else if (device_param->device_vendor_id == VENDOR_ID_NV)
         {
           hm_device_val_to_str ((char *) fanspeed, HM_STR_BUF_SIZE, "%", hm_get_fanspeed_with_device_id (device_id));
         }
 
-        log_info ("HWMon.GPU.#%d...: %s Util, %s Temp, %s Fan", device_id + 1, utilization, temperature, fanspeed);
+        log_info ("HWMon.GPU.#%d...: %s Util, %s Temp, %s Fan, %s Core, %s Memory", device_id + 1, utilization, temperature, fanspeed, corespeed, memoryspeed);
       }
       else
       {
@@ -1598,7 +1586,7 @@ void status_display ()
   #endif // HAVE_HWMON
 }
 
-static void status_benchmark_automat ()
+static void status_benchmark_automate ()
 {
   u64    speed_cnt[DEVICES_MAX] = { 0 };
   double speed_ms[DEVICES_MAX]  = { 0 };
@@ -1645,9 +1633,9 @@ static void status_benchmark ()
   if (data.devices_status == STATUS_STARTING) return;
   if (data.devices_status == STATUS_BYPASS)   return;
 
-  if (data.status_automat == 1)
+  if (data.machine_readable == 1)
   {
-    status_benchmark_automat ();
+    status_benchmark_automate ();
 
     return;
   }
@@ -2215,6 +2203,8 @@ static void check_cracked (hc_device_param_t *device_param, const uint salt_pos)
 
       if (data.digests_shown[hash_pos] == 1) continue;
 
+      hc_thread_mutex_lock (mux_display);
+
       if ((data.opts_type & OPTS_TYPE_PT_NEVERCRACK) == 0)
       {
         data.digests_shown[hash_pos] = 1;
@@ -2235,6 +2225,8 @@ static void check_cracked (hc_device_param_t *device_param, const uint salt_pos)
 
       if (data.salts_done == data.salts_cnt) data.devices_status = STATUS_CRACKED;
 
+      hc_thread_mutex_unlock (mux_display);
+
       check_hash (device_param, &cracked[i]);
     }
 
@@ -2242,6 +2234,8 @@ static void check_cracked (hc_device_param_t *device_param, const uint salt_pos)
 
     if (cpt_cracked > 0)
     {
+      hc_thread_mutex_lock (mux_display);
+
       data.cpt_buf[data.cpt_pos].timestamp = time (NULL);
       data.cpt_buf[data.cpt_pos].cracked   = cpt_cracked;
 
@@ -2250,6 +2244,8 @@ static void check_cracked (hc_device_param_t *device_param, const uint salt_pos)
       data.cpt_total += cpt_cracked;
 
       if (data.cpt_pos == CPT_BUF) data.cpt_pos = 0;
+
+      hc_thread_mutex_unlock (mux_display);
     }
 
     if (data.opts_type & OPTS_TYPE_PT_NEVERCRACK)
@@ -2601,11 +2597,59 @@ static void run_kernel_amp (hc_device_param_t *device_param, const uint num)
   hc_clFinish (data.ocl, device_param->command_queue);
 }
 
+static void run_kernel_memset (hc_device_param_t *device_param, cl_mem buf, const uint value, const uint num)
+{
+  const u32 num16d = num / 16;
+  const u32 num16m = num % 16;
+
+  if (num16d)
+  {
+    device_param->kernel_params_memset_buf32[1] = value;
+    device_param->kernel_params_memset_buf32[2] = num16d;
+
+    uint kernel_threads = device_param->kernel_threads;
+
+    uint num_elements = num16d;
+
+    while (num_elements % kernel_threads) num_elements++;
+
+    cl_kernel kernel = device_param->kernel_memset;
+
+    hc_clSetKernelArg (data.ocl, kernel, 0, sizeof (cl_mem),  (void *) &buf);
+    hc_clSetKernelArg (data.ocl, kernel, 1, sizeof (cl_uint), device_param->kernel_params_memset[1]);
+    hc_clSetKernelArg (data.ocl, kernel, 2, sizeof (cl_uint), device_param->kernel_params_memset[2]);
+
+    const size_t global_work_size[3] = { num_elements,   1, 1 };
+    const size_t local_work_size[3]  = { kernel_threads, 1, 1 };
+
+    hc_clEnqueueNDRangeKernel (data.ocl, device_param->command_queue, kernel, 1, NULL, global_work_size, local_work_size, 0, NULL, NULL);
+
+    hc_clFlush (data.ocl, device_param->command_queue);
+
+    hc_clFinish (data.ocl, device_param->command_queue);
+  }
+
+  if (num16m)
+  {
+    u32 tmp[4];
+
+    tmp[0] = value;
+    tmp[1] = value;
+    tmp[2] = value;
+    tmp[3] = value;
+
+    hc_clEnqueueWriteBuffer (data.ocl, device_param->command_queue, buf, CL_TRUE, num16d * 16, num16m, tmp, 0, NULL, NULL);
+  }
+}
+
 static void run_kernel_bzero (hc_device_param_t *device_param, cl_mem buf, const size_t size)
 {
+  run_kernel_memset (device_param, buf, 0, size);
+
+  /*
   int rc = -1;
 
-  if (device_param->opencl_v12 && device_param->vendor_id == VENDOR_ID_AMD)
+  if (device_param->opencl_v12 && device_param->platform_vendor_id == VENDOR_ID_AMD)
   {
     // So far tested, amd is the only supporting this OpenCL 1.2 function without segfaulting
 
@@ -2636,6 +2680,7 @@ static void run_kernel_bzero (hc_device_param_t *device_param, cl_mem buf, const
 
     myfree (tmp);
   }
+  */
 }
 
 static void choose_kernel (hc_device_param_t *device_param, const uint attack_exec, const uint attack_mode, const uint opts_type, const salt_t *salt_buf, const uint highest_pw_len, const uint pws_cnt)
@@ -2887,6 +2932,14 @@ static void autotune (hc_device_param_t *device_param)
 
   const u32 kernel_power_max = device_param->device_processors * device_param->kernel_threads * kernel_accel_max;
 
+  run_kernel_memset (device_param, device_param->d_pws_buf, 7, kernel_power_max * sizeof (pw_t));
+
+  if (data.attack_exec == ATTACK_EXEC_OUTSIDE_KERNEL)
+  {
+    run_kernel_memset (device_param, device_param->d_pws_amp_buf, 7, kernel_power_max * sizeof (pw_t));
+  }
+
+  /*
   for (u32 i = 0; i < kernel_power_max; i++)
   {
     device_param->pws_buf[i].i[0]   = i;
@@ -2900,6 +2953,7 @@ static void autotune (hc_device_param_t *device_param)
   {
     run_kernel_amp (device_param, kernel_power_max);
   }
+  */
 
   #define VERIFIER_CNT 1
 
@@ -2967,16 +3021,27 @@ static void autotune (hc_device_param_t *device_param)
     exec_ms_pre_final = MIN (exec_ms_pre_final, exec_ms_pre_final_v);
   }
 
+  u32 diff = kernel_loops - kernel_accel;
+
   if ((kernel_loops_min < kernel_loops_max) && (kernel_accel_min < kernel_accel_max))
   {
-    for (u32 f = 2; f < 1024; f++)
+    u32 kernel_accel_orig = kernel_accel;
+    u32 kernel_loops_orig = kernel_loops;
+
+    for (u32 f = 1; f < 1024; f++)
     {
-      const u32 kernel_accel_try = kernel_accel * f;
-      const u32 kernel_loops_try = kernel_loops / f;
+      const u32 kernel_accel_try = (float) kernel_accel_orig * f;
+      const u32 kernel_loops_try = (float) kernel_loops_orig / f;
 
       if (kernel_accel_try > kernel_accel_max) break;
       if (kernel_loops_try < kernel_loops_min) break;
 
+      u32 diff_new = kernel_loops_try - kernel_accel_try;
+
+      if (diff_new > diff) break;
+
+      diff_new = diff;
+
       double exec_ms = try_run (device_param, kernel_accel_try, kernel_loops_try);
 
       for (int i = 0; i < VERIFIER_CNT; i++)
@@ -3011,10 +3076,19 @@ static void autotune (hc_device_param_t *device_param)
 
   // reset them fake words
 
+  /*
   memset (device_param->pws_buf, 0, kernel_power_max * sizeof (pw_t));
 
   hc_clEnqueueWriteBuffer (data.ocl, device_param->command_queue, device_param->d_pws_buf,     CL_TRUE, 0, kernel_power_max * sizeof (pw_t), device_param->pws_buf, 0, NULL, NULL);
   hc_clEnqueueWriteBuffer (data.ocl, device_param->command_queue, device_param->d_pws_amp_buf, CL_TRUE, 0, kernel_power_max * sizeof (pw_t), device_param->pws_buf, 0, NULL, NULL);
+  */
+
+  run_kernel_memset (device_param, device_param->d_pws_buf, 0, kernel_power_max * sizeof (pw_t));
+
+  if (data.attack_exec == ATTACK_EXEC_OUTSIDE_KERNEL)
+  {
+    run_kernel_memset (device_param, device_param->d_pws_amp_buf, 0, kernel_power_max * sizeof (pw_t));
+  }
 
   // reset timer
 
@@ -3311,12 +3385,8 @@ static void run_cracker (hc_device_param_t *device_param, const uint pws_cnt)
        * result
        */
 
-      hc_thread_mutex_lock (mux_display);
-
       check_cracked (device_param, salt_pos);
 
-      hc_thread_mutex_unlock (mux_display);
-
       /**
        * progress
        */
@@ -3339,15 +3409,15 @@ static void run_cracker (hc_device_param_t *device_param, const uint pws_cnt)
 
       hc_timer_set (&device_param->timer_speed);
 
-      hc_thread_mutex_lock (mux_display);
-
       // current speed
 
+      //hc_thread_mutex_lock (mux_display);
+
       device_param->speed_cnt[speed_pos] = perf_sum_all;
 
       device_param->speed_ms[speed_pos] = speed_ms;
 
-      hc_thread_mutex_unlock (mux_display);
+      //hc_thread_mutex_unlock (mux_display);
 
       speed_pos++;
 
@@ -3775,6 +3845,7 @@ static void *thread_monitor (void *p)
 
     if (data.devices_status != STATUS_RUNNING) continue;
 
+
     #ifdef HAVE_HWMON
     if (hwmon_check == 1)
     {
@@ -3923,7 +3994,7 @@ static void *thread_monitor (void *p)
 
       if (status_left == 0)
       {
-        hc_thread_mutex_lock (mux_display);
+        //hc_thread_mutex_lock (mux_display);
 
         if (data.quiet == 0) clear_prompt ();
 
@@ -3933,7 +4004,7 @@ static void *thread_monitor (void *p)
 
         if (data.quiet == 0) log_info ("");
 
-        hc_thread_mutex_unlock (mux_display);
+        //hc_thread_mutex_unlock (mux_display);
 
         status_left = data.status_timer;
       }
@@ -5266,8 +5337,43 @@ static uint generate_bitmaps (const uint digests_cnt, const uint dgst_size, cons
  * main
  */
 
+#ifdef _WIN
+void SetConsoleWindowSize (const int x, const int y)
+{
+  HANDLE h = GetStdHandle (STD_OUTPUT_HANDLE);
+
+  if (h == INVALID_HANDLE_VALUE) return;
+
+  CONSOLE_SCREEN_BUFFER_INFO bufferInfo;
+
+  if (!GetConsoleScreenBufferInfo (h, &bufferInfo)) return;
+
+  SMALL_RECT *sr = &bufferInfo.srWindow;
+
+  sr->Left   = 0;
+  sr->Top    = 0;
+  sr->Right  = MAX (sr->Right,  x - 1);
+  sr->Bottom = MAX (sr->Bottom, y - 1);
+
+  COORD co;
+
+  co.X = sr->Right  + 1;
+  co.Y = sr->Bottom + 1;
+
+  co.Y = MAX (co.Y, 1337);
+
+  if (!SetConsoleScreenBufferSize (h, co)) return;
+
+  if (!SetConsoleWindowInfo (h, TRUE, sr)) return;
+}
+#endif
+
 int main (int argc, char **argv)
 {
+  #ifdef _WIN
+  SetConsoleWindowSize (132, 44);
+  #endif
+
   /**
    * To help users a bit
    */
@@ -5329,87 +5435,87 @@ int main (int argc, char **argv)
    * commandline parameters
    */
 
-  uint  usage                 = USAGE;
-  uint  version               = VERSION;
-  uint  quiet                 = QUIET;
-  uint  benchmark             = BENCHMARK;
-  uint  show                  = SHOW;
-  uint  left                  = LEFT;
-  uint  username              = USERNAME;
-  uint  remove                = REMOVE;
-  uint  remove_timer          = REMOVE_TIMER;
-  u64   skip                  = SKIP;
-  u64   limit                 = LIMIT;
-  uint  keyspace              = KEYSPACE;
-  uint  potfile_disable       = POTFILE_DISABLE;
-  char *potfile_path          = NULL;
-  uint  debug_mode            = DEBUG_MODE;
-  char *debug_file            = NULL;
-  char *induction_dir         = NULL;
-  char *outfile_check_dir     = NULL;
-  uint  force                 = FORCE;
-  uint  runtime               = RUNTIME;
-  uint  hash_mode             = HASH_MODE;
-  uint  attack_mode           = ATTACK_MODE;
-  uint  markov_disable        = MARKOV_DISABLE;
-  uint  markov_classic        = MARKOV_CLASSIC;
-  uint  markov_threshold      = MARKOV_THRESHOLD;
-  char *markov_hcstat         = NULL;
-  char *outfile               = NULL;
-  uint  outfile_format        = OUTFILE_FORMAT;
-  uint  outfile_autohex       = OUTFILE_AUTOHEX;
-  uint  outfile_check_timer   = OUTFILE_CHECK_TIMER;
-  uint  restore               = RESTORE;
-  uint  restore_timer         = RESTORE_TIMER;
-  uint  restore_disable       = RESTORE_DISABLE;
-  uint  status                = STATUS;
-  uint  status_timer          = STATUS_TIMER;
-  uint  status_automat        = STATUS_AUTOMAT;
-  uint  loopback              = LOOPBACK;
-  uint  weak_hash_threshold   = WEAK_HASH_THRESHOLD;
-  char *session               = NULL;
-  uint  hex_charset           = HEX_CHARSET;
-  uint  hex_salt              = HEX_SALT;
-  uint  hex_wordlist          = HEX_WORDLIST;
-  uint  rp_gen                = RP_GEN;
-  uint  rp_gen_func_min       = RP_GEN_FUNC_MIN;
-  uint  rp_gen_func_max       = RP_GEN_FUNC_MAX;
-  uint  rp_gen_seed           = RP_GEN_SEED;
-  char *rule_buf_l            = (char *) RULE_BUF_L;
-  char *rule_buf_r            = (char *) RULE_BUF_R;
-  uint  increment             = INCREMENT;
-  uint  increment_min         = INCREMENT_MIN;
-  uint  increment_max         = INCREMENT_MAX;
-  char *cpu_affinity          = NULL;
-  OCL_PTR *ocl                = NULL;
-  char *opencl_devices        = NULL;
-  char *opencl_platforms      = NULL;
-  char *opencl_device_types   = NULL;
-  uint  opencl_vector_width   = OPENCL_VECTOR_WIDTH;
-  char *truecrypt_keyfiles    = NULL;
-  char *veracrypt_keyfiles    = NULL;
-  uint  veracrypt_pim         = 0;
-  uint  workload_profile      = WORKLOAD_PROFILE;
-  uint  kernel_accel          = KERNEL_ACCEL;
-  uint  kernel_loops          = KERNEL_LOOPS;
-  uint  gpu_temp_disable      = GPU_TEMP_DISABLE;
+  uint  usage                     = USAGE;
+  uint  version                   = VERSION;
+  uint  quiet                     = QUIET;
+  uint  benchmark                 = BENCHMARK;
+  uint  show                      = SHOW;
+  uint  left                      = LEFT;
+  uint  username                  = USERNAME;
+  uint  remove                    = REMOVE;
+  uint  remove_timer              = REMOVE_TIMER;
+  u64   skip                      = SKIP;
+  u64   limit                     = LIMIT;
+  uint  keyspace                  = KEYSPACE;
+  uint  potfile_disable           = POTFILE_DISABLE;
+  char *potfile_path              = NULL;
+  uint  debug_mode                = DEBUG_MODE;
+  char *debug_file                = NULL;
+  char *induction_dir             = NULL;
+  char *outfile_check_dir         = NULL;
+  uint  force                     = FORCE;
+  uint  runtime                   = RUNTIME;
+  uint  hash_mode                 = HASH_MODE;
+  uint  attack_mode               = ATTACK_MODE;
+  uint  markov_disable            = MARKOV_DISABLE;
+  uint  markov_classic            = MARKOV_CLASSIC;
+  uint  markov_threshold          = MARKOV_THRESHOLD;
+  char *markov_hcstat             = NULL;
+  char *outfile                   = NULL;
+  uint  outfile_format            = OUTFILE_FORMAT;
+  uint  outfile_autohex           = OUTFILE_AUTOHEX;
+  uint  outfile_check_timer       = OUTFILE_CHECK_TIMER;
+  uint  restore                   = RESTORE;
+  uint  restore_timer             = RESTORE_TIMER;
+  uint  restore_disable           = RESTORE_DISABLE;
+  uint  status                    = STATUS;
+  uint  status_timer              = STATUS_TIMER;
+  uint  machine_readable          = MACHINE_READABLE;
+  uint  loopback                  = LOOPBACK;
+  uint  weak_hash_threshold       = WEAK_HASH_THRESHOLD;
+  char *session                   = NULL;
+  uint  hex_charset               = HEX_CHARSET;
+  uint  hex_salt                  = HEX_SALT;
+  uint  hex_wordlist              = HEX_WORDLIST;
+  uint  rp_gen                    = RP_GEN;
+  uint  rp_gen_func_min           = RP_GEN_FUNC_MIN;
+  uint  rp_gen_func_max           = RP_GEN_FUNC_MAX;
+  uint  rp_gen_seed               = RP_GEN_SEED;
+  char *rule_buf_l                = (char *) RULE_BUF_L;
+  char *rule_buf_r                = (char *) RULE_BUF_R;
+  uint  increment                 = INCREMENT;
+  uint  increment_min             = INCREMENT_MIN;
+  uint  increment_max             = INCREMENT_MAX;
+  char *cpu_affinity              = NULL;
+  OCL_PTR *ocl                    = NULL;
+  char *opencl_devices            = NULL;
+  char *opencl_platforms          = NULL;
+  char *opencl_device_types       = NULL;
+  uint  opencl_vector_width       = OPENCL_VECTOR_WIDTH;
+  char *truecrypt_keyfiles        = NULL;
+  char *veracrypt_keyfiles        = NULL;
+  uint  veracrypt_pim             = 0;
+  uint  workload_profile          = WORKLOAD_PROFILE;
+  uint  kernel_accel              = KERNEL_ACCEL;
+  uint  kernel_loops              = KERNEL_LOOPS;
+  uint  gpu_temp_disable          = GPU_TEMP_DISABLE;
   #ifdef HAVE_HWMON
-  uint  gpu_temp_abort        = GPU_TEMP_ABORT;
-  uint  gpu_temp_retain       = GPU_TEMP_RETAIN;
+  uint  gpu_temp_abort            = GPU_TEMP_ABORT;
+  uint  gpu_temp_retain           = GPU_TEMP_RETAIN;
   #ifdef HAVE_ADL
-  uint  powertune_enable      = POWERTUNE_ENABLE;
+  uint  powertune_enable          = POWERTUNE_ENABLE;
   #endif
   #endif
-  uint  logfile_disable       = LOGFILE_DISABLE;
-  uint  segment_size          = SEGMENT_SIZE;
-  uint  scrypt_tmto           = SCRYPT_TMTO;
-  char  separator             = SEPARATOR;
-  uint  bitmap_min            = BITMAP_MIN;
-  uint  bitmap_max            = BITMAP_MAX;
-  char *custom_charset_1      = NULL;
-  char *custom_charset_2      = NULL;
-  char *custom_charset_3      = NULL;
-  char *custom_charset_4      = NULL;
+  uint  logfile_disable           = LOGFILE_DISABLE;
+  uint  segment_size              = SEGMENT_SIZE;
+  uint  scrypt_tmto               = SCRYPT_TMTO;
+  char  separator                 = SEPARATOR;
+  uint  bitmap_min                = BITMAP_MIN;
+  uint  bitmap_max                = BITMAP_MAX;
+  char *custom_charset_1          = NULL;
+  char *custom_charset_2          = NULL;
+  char *custom_charset_3          = NULL;
+  char *custom_charset_4          = NULL;
 
   #define IDX_HELP                      'h'
   #define IDX_VERSION                   'V'
@@ -5452,7 +5558,7 @@ int main (int argc, char **argv)
   #define IDX_RESTORE_DISABLE           0xff27
   #define IDX_STATUS                    0xff17
   #define IDX_STATUS_TIMER              0xff18
-  #define IDX_STATUS_AUTOMAT            0xff50
+  #define IDX_MACHINE_READABLE          0xff50
   #define IDX_LOOPBACK                  0xff38
   #define IDX_WEAK_HASH_THRESHOLD       0xff42
   #define IDX_SESSION                   0xff19
@@ -5516,7 +5622,7 @@ int main (int argc, char **argv)
     {"restore-disable",           no_argument,       0, IDX_RESTORE_DISABLE},
     {"status",                    no_argument,       0, IDX_STATUS},
     {"status-timer",              required_argument, 0, IDX_STATUS_TIMER},
-    {"status-automat",            no_argument,       0, IDX_STATUS_AUTOMAT},
+    {"machine-readable",          no_argument,       0, IDX_MACHINE_READABLE},
     {"loopback",                  no_argument,       0, IDX_LOOPBACK},
     {"weak-hash-threshold",       required_argument, 0, IDX_WEAK_HASH_THRESHOLD},
     {"session",                   required_argument, 0, IDX_SESSION},
@@ -5819,7 +5925,7 @@ int main (int argc, char **argv)
       case IDX_RESTORE_DISABLE:           restore_disable           = 1;              break;
       case IDX_STATUS:                    status                    = 1;              break;
       case IDX_STATUS_TIMER:              status_timer              = atoi (optarg);  break;
-      case IDX_STATUS_AUTOMAT:            status_automat            = 1;              break;
+      case IDX_MACHINE_READABLE:          machine_readable          = 1;              break;
       case IDX_LOOPBACK:                  loopback                  = 1;              break;
       case IDX_WEAK_HASH_THRESHOLD:       weak_hash_threshold       = atoi (optarg);  break;
     //case IDX_SESSION:                   session                   = optarg;         break;
@@ -5920,7 +6026,7 @@ int main (int argc, char **argv)
   {
     if (benchmark == 1)
     {
-      if (status_automat == 0)
+      if (machine_readable == 0)
       {
         log_info ("%s (%s) starting in benchmark-mode...", PROGNAME, VERSION_TAG);
         log_info ("");
@@ -6550,47 +6656,47 @@ int main (int argc, char **argv)
    * store stuff
    */
 
-  data.hash_mode          = hash_mode;
-  data.restore            = restore;
-  data.restore_timer      = restore_timer;
-  data.restore_disable    = restore_disable;
-  data.status             = status;
-  data.status_timer       = status_timer;
-  data.status_automat     = status_automat;
-  data.loopback           = loopback;
-  data.runtime            = runtime;
-  data.remove             = remove;
-  data.remove_timer       = remove_timer;
-  data.debug_mode         = debug_mode;
-  data.debug_file         = debug_file;
-  data.username           = username;
-  data.quiet              = quiet;
-  data.outfile            = outfile;
-  data.outfile_format     = outfile_format;
-  data.outfile_autohex    = outfile_autohex;
-  data.hex_charset        = hex_charset;
-  data.hex_salt           = hex_salt;
-  data.hex_wordlist       = hex_wordlist;
-  data.separator          = separator;
-  data.rp_files           = rp_files;
-  data.rp_files_cnt       = rp_files_cnt;
-  data.rp_gen             = rp_gen;
-  data.rp_gen_seed        = rp_gen_seed;
-  data.force              = force;
-  data.benchmark          = benchmark;
-  data.skip               = skip;
-  data.limit              = limit;
+  data.hash_mode               = hash_mode;
+  data.restore                 = restore;
+  data.restore_timer           = restore_timer;
+  data.restore_disable         = restore_disable;
+  data.status                  = status;
+  data.status_timer            = status_timer;
+  data.machine_readable        = machine_readable;
+  data.loopback                = loopback;
+  data.runtime                 = runtime;
+  data.remove                  = remove;
+  data.remove_timer            = remove_timer;
+  data.debug_mode              = debug_mode;
+  data.debug_file              = debug_file;
+  data.username                = username;
+  data.quiet                   = quiet;
+  data.outfile                 = outfile;
+  data.outfile_format          = outfile_format;
+  data.outfile_autohex         = outfile_autohex;
+  data.hex_charset             = hex_charset;
+  data.hex_salt                = hex_salt;
+  data.hex_wordlist            = hex_wordlist;
+  data.separator               = separator;
+  data.rp_files                = rp_files;
+  data.rp_files_cnt            = rp_files_cnt;
+  data.rp_gen                  = rp_gen;
+  data.rp_gen_seed             = rp_gen_seed;
+  data.force                   = force;
+  data.benchmark               = benchmark;
+  data.skip                    = skip;
+  data.limit                   = limit;
   #ifdef HAVE_HWMON
   #ifdef HAVE_ADL
-  data.powertune_enable   = powertune_enable;
+  data.powertune_enable        = powertune_enable;
   #endif
   #endif
-  data.logfile_disable    = logfile_disable;
-  data.truecrypt_keyfiles = truecrypt_keyfiles;
-  data.veracrypt_keyfiles = veracrypt_keyfiles;
-  data.veracrypt_pim      = veracrypt_pim;
-  data.scrypt_tmto        = scrypt_tmto;
-  data.workload_profile   = workload_profile;
+  data.logfile_disable         = logfile_disable;
+  data.truecrypt_keyfiles      = truecrypt_keyfiles;
+  data.veracrypt_keyfiles      = veracrypt_keyfiles;
+  data.veracrypt_pim           = veracrypt_pim;
+  data.scrypt_tmto             = scrypt_tmto;
+  data.workload_profile        = workload_profile;
 
   /**
    * cpu affinity
@@ -6704,7 +6810,7 @@ int main (int argc, char **argv)
   logfile_top_uint   (segment_size);
   logfile_top_uint   (show);
   logfile_top_uint   (status);
-  logfile_top_uint   (status_automat);
+  logfile_top_uint   (machine_readable);
   logfile_top_uint   (status_timer);
   logfile_top_uint   (usage);
   logfile_top_uint   (username);
@@ -6782,6 +6888,7 @@ int main (int argc, char **argv)
     potfile_disable       = 1;
     weak_hash_threshold   = 0;
     gpu_temp_disable      = 1;
+    powertune_enable      = 1;
 
     data.status_timer     = status_timer;
     data.restore_timer    = restore_timer;
@@ -11090,6 +11197,8 @@ int main (int argc, char **argv)
                   break;
       case  7400: if (pw_max > 16) pw_max = 16;
                   break;
+      case  7500: if (pw_max >  8) pw_max =  8;
+                  break;
       case  7900: if (pw_max > 48) pw_max = 48;
                   break;
       case  8500: if (pw_max >  8) pw_max =  8;
@@ -13201,39 +13310,43 @@ int main (int argc, char **argv)
       // this causes trouble with vendor id based macros
       // we'll assign generic to those without special optimization available
 
-      cl_uint vendor_id = 0;
+      cl_uint platform_vendor_id = 0;
 
       if (strcmp (platform_vendor, CL_VENDOR_AMD) == 0)
       {
-        vendor_id = VENDOR_ID_AMD;
+        platform_vendor_id = VENDOR_ID_AMD;
+      }
+      else if (strcmp (platform_vendor, CL_VENDOR_AMD_USE_INTEL) == 0)
+      {
+        platform_vendor_id = VENDOR_ID_AMD_USE_INTEL;
       }
       else if (strcmp (platform_vendor, CL_VENDOR_APPLE) == 0)
       {
-        vendor_id = VENDOR_ID_APPLE;
+        platform_vendor_id = VENDOR_ID_APPLE;
       }
       else if (strcmp (platform_vendor, CL_VENDOR_INTEL_BEIGNET) == 0)
       {
-        vendor_id = VENDOR_ID_INTEL_BEIGNET;
+        platform_vendor_id = VENDOR_ID_INTEL_BEIGNET;
       }
       else if (strcmp (platform_vendor, CL_VENDOR_INTEL_SDK) == 0)
       {
-        vendor_id = VENDOR_ID_INTEL_SDK;
+        platform_vendor_id = VENDOR_ID_INTEL_SDK;
       }
       else if (strcmp (platform_vendor, CL_VENDOR_MESA) == 0)
       {
-        vendor_id = VENDOR_ID_MESA;
+        platform_vendor_id = VENDOR_ID_MESA;
       }
       else if (strcmp (platform_vendor, CL_VENDOR_NV) == 0)
       {
-        vendor_id = VENDOR_ID_NV;
+        platform_vendor_id = VENDOR_ID_NV;
       }
       else if (strcmp (platform_vendor, CL_VENDOR_POCL) == 0)
       {
-        vendor_id = VENDOR_ID_POCL;
+        platform_vendor_id = VENDOR_ID_POCL;
       }
       else
       {
-        vendor_id = VENDOR_ID_GENERIC;
+        platform_vendor_id = VENDOR_ID_GENERIC;
       }
 
       for (uint platform_devices_id = 0; platform_devices_id < platform_devices_cnt; platform_devices_id++)
@@ -13244,7 +13357,7 @@ int main (int argc, char **argv)
 
         hc_device_param_t *device_param = &data.devices_param[device_id];
 
-        device_param->vendor_id = vendor_id;
+        device_param->platform_vendor_id = platform_vendor_id;
 
         device_param->device = platform_devices[platform_devices_id];
 
@@ -13272,6 +13385,57 @@ int main (int argc, char **argv)
 
         device_param->device_name = device_name;
 
+        // device_vendor
+
+        hc_clGetDeviceInfo (data.ocl, device_param->device, CL_DEVICE_VENDOR, 0, NULL, &param_value_size);
+
+        char *device_vendor = (char *) mymalloc (param_value_size);
+
+        hc_clGetDeviceInfo (data.ocl, device_param->device, CL_DEVICE_VENDOR, param_value_size, device_vendor, NULL);
+
+        device_param->device_vendor = device_vendor;
+
+        cl_uint device_vendor_id = 0;
+
+        if (strcmp (device_vendor, CL_VENDOR_AMD) == 0)
+        {
+          device_vendor_id = VENDOR_ID_AMD;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_AMD_USE_INTEL) == 0)
+        {
+          device_vendor_id = VENDOR_ID_AMD_USE_INTEL;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_APPLE) == 0)
+        {
+          device_vendor_id = VENDOR_ID_APPLE;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_INTEL_BEIGNET) == 0)
+        {
+          device_vendor_id = VENDOR_ID_INTEL_BEIGNET;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_INTEL_SDK) == 0)
+        {
+          device_vendor_id = VENDOR_ID_INTEL_SDK;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_MESA) == 0)
+        {
+          device_vendor_id = VENDOR_ID_MESA;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_NV) == 0)
+        {
+          device_vendor_id = VENDOR_ID_NV;
+        }
+        else if (strcmp (device_vendor, CL_VENDOR_POCL) == 0)
+        {
+          device_vendor_id = VENDOR_ID_POCL;
+        }
+        else
+        {
+          device_vendor_id = VENDOR_ID_GENERIC;
+        }
+
+        device_param->device_vendor_id = device_vendor_id;
+
         // tuning db
 
         tuning_db_entry_t *tuningdb_entry = tuning_db_search (tuning_db, device_param, attack_mode, hash_mode);
@@ -13378,7 +13542,7 @@ int main (int argc, char **argv)
 
         if (device_endian_little == CL_FALSE)
         {
-          log_info ("Device #%u: WARNING: not little endian device", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: not little endian device", device_id + 1);
 
           device_param->skipped = 1;
         }
@@ -13391,7 +13555,7 @@ int main (int argc, char **argv)
 
         if (device_available == CL_FALSE)
         {
-          log_info ("Device #%u: WARNING: device not available", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: device not available", device_id + 1);
 
           device_param->skipped = 1;
         }
@@ -13404,7 +13568,7 @@ int main (int argc, char **argv)
 
         if (device_compiler_available == CL_FALSE)
         {
-          log_info ("Device #%u: WARNING: device no compiler available", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: device no compiler available", device_id + 1);
 
           device_param->skipped = 1;
         }
@@ -13417,7 +13581,7 @@ int main (int argc, char **argv)
 
         if ((device_execution_capabilities & CL_EXEC_KERNEL) == 0)
         {
-          log_info ("Device #%u: WARNING: device does not support executing kernels", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: device does not support executing kernels", device_id + 1);
 
           device_param->skipped = 1;
         }
@@ -13434,14 +13598,14 @@ int main (int argc, char **argv)
 
         if (strstr (device_extensions, "base_atomics") == 0)
         {
-          log_info ("Device #%u: WARNING: device does not support base atomics", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: device does not support base atomics", device_id + 1);
 
           device_param->skipped = 1;
         }
 
         if (strstr (device_extensions, "byte_addressable_store") == 0)
         {
-          log_info ("Device #%u: WARNING: device does not support byte addressable store", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: device does not support byte addressable store", device_id + 1);
 
           device_param->skipped = 1;
         }
@@ -13456,11 +13620,26 @@ int main (int argc, char **argv)
 
         if (device_local_mem_size < 32768)
         {
-          log_info ("Device #%u: WARNING: device local mem size is too small", device_id + 1);
+          if (data.quiet == 0) log_info ("Device #%u: WARNING: device local mem size is too small", device_id + 1);
 
           device_param->skipped = 1;
         }
 
+        // If there's both an Intel CPU and an AMD OpenCL runtime it's a tricky situation
+        // Both platforms support CPU device types and therefore both will try to use 100% of the physical resources
+        // This results in both utilizing it for 50%
+        // However, Intel has much better SIMD control over their own hardware
+        // It makes sense to give them full control over their own hardware
+
+        if (device_type & CL_DEVICE_TYPE_CPU)
+        {
+          if (device_param->device_vendor_id == VENDOR_ID_AMD_USE_INTEL)
+          {
+            if (data.quiet == 0) log_info ("Device #%u: WARNING: not native intel opencl platform", device_id + 1);
+
+            device_param->skipped = 1;
+          }
+        }
 
         // skipped
 
@@ -13482,9 +13661,9 @@ int main (int argc, char **argv)
         char *device_name_chksum = (char *) mymalloc (INFOSZ);
 
         #if __x86_64__
-        snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 64, device_param->vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
+        snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 64, device_param->platform_vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
         #else
-        snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 32, device_param->vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
+        snprintf (device_name_chksum, INFOSZ - 1, "%u-%u-%u-%s-%s-%s-%u", 32, device_param->platform_vendor_id, device_param->vector_width, device_param->device_name, device_param->device_version, device_param->driver_version, COMPTIME);
         #endif
 
         uint device_name_digest[4] = { 0 };
@@ -13506,7 +13685,7 @@ int main (int argc, char **argv)
 
         if (device_type & CL_DEVICE_TYPE_GPU)
         {
-          if (vendor_id == VENDOR_ID_AMD)
+          if (device_vendor_id == VENDOR_ID_AMD)
           {
             cl_uint device_processor_cores = 0;
 
@@ -13516,7 +13695,7 @@ int main (int argc, char **argv)
 
             device_param->device_processor_cores = device_processor_cores;
           }
-          else if (vendor_id == VENDOR_ID_NV)
+          else if (device_vendor_id == VENDOR_ID_NV)
           {
             cl_uint kernel_exec_timeout = 0;
 
@@ -13558,7 +13737,7 @@ int main (int argc, char **argv)
 
         if ((benchmark == 1 || quiet == 0) && (algorithm_pos == 0))
         {
-          if (status_automat == 0)
+          if (machine_readable == 0)
           {
             if (device_param->skipped == 0)
             {
@@ -13585,7 +13764,7 @@ int main (int argc, char **argv)
         {
           if (device_type & CL_DEVICE_TYPE_GPU)
           {
-            if (vendor_id == VENDOR_ID_AMD)
+            if (platform_vendor_id == VENDOR_ID_AMD)
             {
               int catalyst_check = (force == 1) ? 0 : 1;
 
@@ -13633,7 +13812,7 @@ int main (int argc, char **argv)
                 return (-1);
               }
             }
-            else if (vendor_id == VENDOR_ID_NV)
+            else if (platform_vendor_id == VENDOR_ID_NV)
             {
               if (device_param->kernel_exec_timeout != 0)
               {
@@ -13643,9 +13822,10 @@ int main (int argc, char **argv)
             }
           }
 
+          /* turns out pocl still creates segfaults (because of llvm)
           if (device_type & CL_DEVICE_TYPE_CPU)
           {
-            if (vendor_id == VENDOR_ID_AMD)
+            if (platform_vendor_id == VENDOR_ID_AMD)
             {
               if (force == 0)
               {
@@ -13660,6 +13840,7 @@ int main (int argc, char **argv)
               }
             }
           }
+          */
 
           /**
            * kernel accel and loops tuning db adjustment
@@ -13754,7 +13935,7 @@ int main (int argc, char **argv)
 
     if ((benchmark == 1 || quiet == 0) && (algorithm_pos == 0))
     {
-      if (status_automat == 0)
+      if (machine_readable == 0)
       {
         log_info ("");
       }
@@ -14024,14 +14205,14 @@ int main (int argc, char **argv)
         const uint platform_devices_id = device_param->platform_devices_id;
 
         #if defined(HAVE_NVML) || defined(HAVE_NVAPI)
-        if (device_param->vendor_id == VENDOR_ID_NV)
+        if (device_param->device_vendor_id == VENDOR_ID_NV)
         {
           memcpy (&data.hm_device[device_id], &hm_adapters_nv[platform_devices_id], sizeof (hm_attrs_t));
         }
         #endif
 
         #ifdef HAVE_ADL
-        if (device_param->vendor_id == VENDOR_ID_AMD)
+        if (device_param->device_vendor_id == VENDOR_ID_AMD)
         {
           memcpy (&data.hm_device[device_id], &hm_adapters_amd[platform_devices_id], sizeof (hm_attrs_t));
         }
@@ -14219,22 +14400,22 @@ int main (int argc, char **argv)
 
           if (hash_mode == 8900)
           {
-            if (device_param->vendor_id == VENDOR_ID_AMD)
+            if (device_param->device_vendor_id == VENDOR_ID_AMD)
             {
               tmto_start = 1;
             }
-            else if (device_param->vendor_id == VENDOR_ID_NV)
+            else if (device_param->device_vendor_id == VENDOR_ID_NV)
             {
               tmto_start = 2;
             }
           }
           else if (hash_mode == 9300)
           {
-            if (device_param->vendor_id == VENDOR_ID_AMD)
+            if (device_param->device_vendor_id == VENDOR_ID_AMD)
             {
               tmto_start = 2;
             }
-            else if (device_param->vendor_id == VENDOR_ID_NV)
+            else if (device_param->device_vendor_id == VENDOR_ID_NV)
             {
               tmto_start = 2;
             }
@@ -14465,36 +14646,38 @@ int main (int argc, char **argv)
 
         int skip = 0;
 
-        if (size_pws   > device_param->device_maxmem_alloc) skip = 1;
-        if (size_tmps  > device_param->device_maxmem_alloc) skip = 1;
-        if (size_hooks > device_param->device_maxmem_alloc) skip = 1;
-
-        if (( bitmap_size
-            + bitmap_size
-            + bitmap_size
-            + bitmap_size
-            + bitmap_size
-            + bitmap_size
-            + bitmap_size
-            + bitmap_size
-            + size_bfs
-            + size_combs
-            + size_digests
-            + size_esalts
-            + size_hooks
-            + size_markov_css
-            + size_plains
-            + size_pws
-            + size_pws // not a bug
-            + size_results
-            + size_root_css
-            + size_rules
-            + size_rules_c
-            + size_salts
-            + size_scryptV
-            + size_shown
-            + size_tm
-            + size_tmps) > device_param->device_global_mem) skip = 1;
+        const u64 size_total
+          = bitmap_size
+          + bitmap_size
+          + bitmap_size
+          + bitmap_size
+          + bitmap_size
+          + bitmap_size
+          + bitmap_size
+          + bitmap_size
+          + size_bfs
+          + size_combs
+          + size_digests
+          + size_esalts
+          + size_hooks
+          + size_markov_css
+          + size_plains
+          + size_pws
+          + size_pws // not a bug
+          + size_results
+          + size_root_css
+          + size_rules
+          + size_rules_c
+          + size_salts
+          + size_scryptV
+          + size_shown
+          + size_tm
+          + size_tmps;
+
+        // Don't ask me, ask AMD!
+
+        if (size_total > device_param->device_maxmem_alloc) skip = 1;
+        if (size_total > device_param->device_global_mem)   skip = 1;
 
         if (skip == 1)
         {
@@ -14552,18 +14735,28 @@ int main (int argc, char **argv)
 
       // we don't have sm_* on vendors not NV but it doesn't matter
 
-      snprintf (build_opts, sizeof (build_opts) - 1, "-cl-std=CL1.1 -I\"%s/\" -DVENDOR_ID=%u -DCUDA_ARCH=%d -DVECT_SIZE=%u -DDEVICE_TYPE=%u -DKERN_TYPE=%u -D_unroll", shared_dir, device_param->vendor_id, (device_param->sm_major * 100) + device_param->sm_minor, device_param->vector_width, (u32) device_param->device_type, kern_type);
+      #if _WIN
+      snprintf (build_opts, sizeof (build_opts) - 1, "-I \"%s\\OpenCL\\\" -I '%s\\OpenCL\\' -I %s\\OpenCL\\ -I\"%s\\OpenCL\\\" -I'%s\\OpenCL\\' -I%s\\OpenCL\\", shared_dir, shared_dir, shared_dir, shared_dir, shared_dir, shared_dir);
+      #else
+      snprintf (build_opts, sizeof (build_opts) - 1, "-I \"%s/OpenCL/\" -I '%s/OpenCL/' -I %s/OpenCL/ -I\"%s/OpenCL/\" -I'%s/OpenCL/' -I%s/OpenCL/", shared_dir, shared_dir, shared_dir, shared_dir, shared_dir, shared_dir);
+      #endif
+
+      char build_opts_new[1024] = { 0 };
 
-      if (device_param->vendor_id == VENDOR_ID_INTEL_SDK)
+      snprintf (build_opts_new, sizeof (build_opts_new) - 1, "%s -DVENDOR_ID=%u -DCUDA_ARCH=%d -DVECT_SIZE=%u -DDEVICE_TYPE=%u -DKERN_TYPE=%u -D_unroll -cl-std=CL1.1", build_opts, device_param->device_vendor_id, (device_param->sm_major * 100) + device_param->sm_minor, device_param->vector_width, (u32) device_param->device_type, kern_type);
+
+      strncpy (build_opts, build_opts_new, sizeof (build_opts) - 1);
+
+      /*
+      if (device_param->device_vendor_id == VENDOR_ID_INTEL_SDK)
       {
         // we do vectorizing much better than the auto-vectorizer
 
-        char build_opts_new[1024] = { 0 };
-
         snprintf (build_opts_new, sizeof (build_opts_new) - 1, "%s -cl-opt-disable", build_opts);
 
         strncpy (build_opts, build_opts_new, sizeof (build_opts) - 1);
       }
+      */
 
       #ifdef DEBUG
       log_info ("Device #%u: build_opts '%s'\n", device_id + 1, build_opts);
@@ -14985,13 +15178,6 @@ int main (int argc, char **argv)
       hc_clEnqueueWriteBuffer (data.ocl, device_param->command_queue, device_param->d_digests_shown,  CL_TRUE, 0, size_shown,   data.digests_shown, 0, NULL, NULL);
       hc_clEnqueueWriteBuffer (data.ocl, device_param->command_queue, device_param->d_salt_bufs,      CL_TRUE, 0, size_salts,   data.salts_buf,     0, NULL, NULL);
 
-      run_kernel_bzero (device_param, device_param->d_pws_buf,        size_pws);
-      run_kernel_bzero (device_param, device_param->d_pws_amp_buf,    size_pws);
-      run_kernel_bzero (device_param, device_param->d_tmps,           size_tmps);
-      run_kernel_bzero (device_param, device_param->d_hooks,          size_hooks);
-      run_kernel_bzero (device_param, device_param->d_plain_bufs,     size_plains);
-      run_kernel_bzero (device_param, device_param->d_result,         size_results);
-
       /**
        * special buffers
        */
@@ -15002,8 +15188,6 @@ int main (int argc, char **argv)
         device_param->d_rules_c = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_rules_c, NULL);
 
         hc_clEnqueueWriteBuffer (data.ocl, device_param->command_queue, device_param->d_rules, CL_TRUE, 0, size_rules, kernel_rules_buf, 0, NULL, NULL);
-
-        run_kernel_bzero (device_param, device_param->d_rules_c, size_rules_c);
       }
       else if (attack_kern == ATTACK_KERN_COMBI)
       {
@@ -15011,11 +15195,6 @@ int main (int argc, char **argv)
         device_param->d_combs_c         = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_combs,      NULL);
         device_param->d_root_css_buf    = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_root_css,   NULL);
         device_param->d_markov_css_buf  = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_markov_css, NULL);
-
-        run_kernel_bzero (device_param, device_param->d_combs,          size_combs);
-        run_kernel_bzero (device_param, device_param->d_combs_c,        size_combs);
-        run_kernel_bzero (device_param, device_param->d_root_css_buf,   size_root_css);
-        run_kernel_bzero (device_param, device_param->d_markov_css_buf, size_markov_css);
       }
       else if (attack_kern == ATTACK_KERN_BF)
       {
@@ -15024,12 +15203,6 @@ int main (int argc, char **argv)
         device_param->d_tm_c            = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_tm,         NULL);
         device_param->d_root_css_buf    = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_root_css,   NULL);
         device_param->d_markov_css_buf  = hc_clCreateBuffer (data.ocl, device_param->context, CL_MEM_READ_ONLY, size_markov_css, NULL);
-
-        run_kernel_bzero (device_param, device_param->d_bfs,            size_bfs);
-        run_kernel_bzero (device_param, device_param->d_bfs_c,          size_bfs);
-        run_kernel_bzero (device_param, device_param->d_tm_c,           size_tm);
-        run_kernel_bzero (device_param, device_param->d_root_css_buf,   size_root_css);
-        run_kernel_bzero (device_param, device_param->d_markov_css_buf, size_markov_css);
       }
 
       if (size_esalts)
@@ -15173,6 +15346,13 @@ int main (int argc, char **argv)
       device_param->kernel_params_tm[0] = &device_param->d_bfs_c;
       device_param->kernel_params_tm[1] = &device_param->d_tm_c;
 
+      device_param->kernel_params_memset_buf32[1] = 0; // value
+      device_param->kernel_params_memset_buf32[2] = 0; // gid_max
+
+      device_param->kernel_params_memset[0] = NULL;
+      device_param->kernel_params_memset[1] = &device_param->kernel_params_memset_buf32[1];
+      device_param->kernel_params_memset[2] = &device_param->kernel_params_memset_buf32[2];
+
       /**
        * kernel name
        */
@@ -15281,6 +15461,18 @@ int main (int argc, char **argv)
         if (opts_type & OPTS_TYPE_HOOK23) hc_clSetKernelArg (data.ocl, device_param->kernel23, i, sizeof (cl_uint), device_param->kernel_params[i]);
       }
 
+      // GPU memset
+
+      device_param->kernel_memset = hc_clCreateKernel (data.ocl, device_param->program, "gpu_memset");
+
+      hc_clGetKernelWorkGroupInfo (data.ocl, device_param->kernel_memset, device_param->device, CL_KERNEL_WORK_GROUP_SIZE, sizeof (size_t), &kernel_wgs_tmp, NULL); kernel_threads = MIN (kernel_threads, kernel_wgs_tmp);
+
+      hc_clSetKernelArg (data.ocl, device_param->kernel_memset, 0, sizeof (cl_mem),  device_param->kernel_params_memset[0]);
+      hc_clSetKernelArg (data.ocl, device_param->kernel_memset, 1, sizeof (cl_uint), device_param->kernel_params_memset[1]);
+      hc_clSetKernelArg (data.ocl, device_param->kernel_memset, 2, sizeof (cl_uint), device_param->kernel_params_memset[2]);
+
+      // MP start
+
       if (attack_mode == ATTACK_MODE_BF)
       {
         device_param->kernel_mp_l = hc_clCreateKernel (data.ocl, device_param->program_mp, "l_markov");
@@ -15341,6 +15533,39 @@ int main (int argc, char **argv)
 
       device_param->kernel_threads = kernel_threads;
 
+      // zero some data buffers
+
+      run_kernel_bzero (device_param, device_param->d_pws_buf,        size_pws);
+      run_kernel_bzero (device_param, device_param->d_pws_amp_buf,    size_pws);
+      run_kernel_bzero (device_param, device_param->d_tmps,           size_tmps);
+      run_kernel_bzero (device_param, device_param->d_hooks,          size_hooks);
+      run_kernel_bzero (device_param, device_param->d_plain_bufs,     size_plains);
+      run_kernel_bzero (device_param, device_param->d_result,         size_results);
+
+      /**
+       * special buffers
+       */
+
+      if (attack_kern == ATTACK_KERN_STRAIGHT)
+      {
+        run_kernel_bzero (device_param, device_param->d_rules_c, size_rules_c);
+      }
+      else if (attack_kern == ATTACK_KERN_COMBI)
+      {
+        run_kernel_bzero (device_param, device_param->d_combs,          size_combs);
+        run_kernel_bzero (device_param, device_param->d_combs_c,        size_combs);
+        run_kernel_bzero (device_param, device_param->d_root_css_buf,   size_root_css);
+        run_kernel_bzero (device_param, device_param->d_markov_css_buf, size_markov_css);
+      }
+      else if (attack_kern == ATTACK_KERN_BF)
+      {
+        run_kernel_bzero (device_param, device_param->d_bfs,            size_bfs);
+        run_kernel_bzero (device_param, device_param->d_bfs_c,          size_bfs);
+        run_kernel_bzero (device_param, device_param->d_tm_c,           size_tm);
+        run_kernel_bzero (device_param, device_param->d_root_css_buf,   size_root_css);
+        run_kernel_bzero (device_param, device_param->d_markov_css_buf, size_markov_css);
+      }
+
       /**
        * Store initial fanspeed if gpu_temp_retain is enabled
        */
@@ -15529,7 +15754,7 @@ int main (int argc, char **argv)
 
     if (benchmark == 1)
     {
-      if (status_automat == 0)
+      if (machine_readable == 0)
       {
         quiet = 0;
 
@@ -17533,7 +17758,7 @@ int main (int argc, char **argv)
     {
       status_benchmark ();
 
-      if (status_automat == 0)
+      if (machine_readable == 0)
       {
         log_info ("");
       }
@@ -17614,6 +17839,7 @@ int main (int argc, char **argv)
       if (device_param->kernel_mp_r)        hc_clReleaseKernel        (data.ocl, device_param->kernel_mp_r);
       if (device_param->kernel_tm)          hc_clReleaseKernel        (data.ocl, device_param->kernel_tm);
       if (device_param->kernel_amp)         hc_clReleaseKernel        (data.ocl, device_param->kernel_amp);
+      if (device_param->kernel_memset)      hc_clReleaseKernel        (data.ocl, device_param->kernel_memset);
 
       if (device_param->program)            hc_clReleaseProgram       (data.ocl, device_param->program);
       if (device_param->program_mp)         hc_clReleaseProgram       (data.ocl, device_param->program_mp);