GOSTSec/ccminer-gostd-lite
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

nvml: small changes to handle the 750ti app clocks

Browse Source 
dynamically alloc the hardcoded gpu clocks list array size.

Unlike the 9xx, the 750ti api call requires to get them all (361.62)

Something was "fixed" recently for the 750 Ti in linux drivers...
This commit is contained in:
Tanguy Pruvot 2016-06-19 20:08:38 +02:00
parent 41a68f8f78
commit fe99de24a8
1 changed files with 37 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

65
nvml.cpp
View File

@ -329,31 +329,36 @@ int nvml_set_clocks(nvml_handle *nvmlh, int dev_id)
if (device_mem_clocks[dev_id]) mem_clk = device_mem_clocks[dev_id]; if (device_mem_clocks[dev_id]) mem_clk = device_mem_clocks[dev_id];
if (device_gpu_clocks[dev_id]) gpu_clk = device_gpu_clocks[dev_id]; if (device_gpu_clocks[dev_id]) gpu_clk = device_gpu_clocks[dev_id];
// these functions works for the 960 and the 970 (346.72+), not for the 750 Ti // these functions works for the 960 and the 970 (346.72+), and for the 750 Ti with driver ~361+
uint32_t nclocks = 0, clocks[127] = { 0 }; uint32_t nclocks = 0, mem_clocks[32] = { 0 };
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, NULL); nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, NULL);
nclocks = min(nclocks, 127); nclocks = min(nclocks, 32);
if (nclocks) if (nclocks)
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, clocks); nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, mem_clocks);
for (uint8_t u=0; u < nclocks; u++) { for (uint8_t u=0; u < nclocks; u++) {
// ordered by pstate (so highest is first memory clock - P0) // ordered by pstate (so highest is first memory clock - P0)
if (clocks[u] <= mem_clk) { if (mem_clocks[u] <= mem_clk) {
mem_clk = clocks[u]; mem_clk = mem_clocks[u];
break; break;
} }
} }
uint32_t* gpu_clocks = NULL;
nclocks = 0; nclocks = 0;
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, NULL); nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, NULL);
nclocks = min(nclocks, 127); if (nclocks) {
if (nclocks) if (opt_debug)
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, clocks); applog(LOG_DEBUG, "GPU #%d: %u clocks found for mem %u", dev_id, nclocks, mem_clk);
for (uint8_t u=0; u < nclocks; u++) { gpu_clocks = (uint32_t*) calloc(1, sizeof(uint32_t) * nclocks + 4);
// ordered desc, so get first nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, gpu_clocks);
if (clocks[u] <= gpu_clk) { for (uint8_t u=0; u < nclocks; u++) {
gpu_clk = clocks[u]; // ordered desc, so get first
break; if (gpu_clocks[u] <= gpu_clk) {
gpu_clk = gpu_clocks[u];
break;
}
} }
free(gpu_clocks);
} }
rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[n], mem_clk, gpu_clk); rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[n], mem_clk, gpu_clk);
@ -435,37 +440,41 @@ int nvml_set_pstate(nvml_handle *nvmlh, int dev_id)
if (device_mem_clocks[dev_id]) mem_clk = device_mem_clocks[dev_id]; if (device_mem_clocks[dev_id]) mem_clk = device_mem_clocks[dev_id];
if (device_gpu_clocks[dev_id]) gpu_clk = device_gpu_clocks[dev_id]; if (device_gpu_clocks[dev_id]) gpu_clk = device_gpu_clocks[dev_id];
// these functions works for the 960 and the 970 (346.72+), not for the 750 Ti // these functions works for the 960 and the 970 (346.72+), and for the 750 Ti with driver ~361+
uint32_t nclocks = 0, clocks[127] = { 0 }; uint32_t nclocks = 0, mem_clocks[32] = { 0 };
int8_t wanted_pstate = device_pstate[dev_id]; int8_t wanted_pstate = device_pstate[dev_id];
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, NULL); nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, NULL);
nclocks = min(nclocks, 127); nclocks = min(nclocks, 32);
if (nclocks) if (nclocks)
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, clocks); nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, mem_clocks);
if (wanted_pstate+1 > nclocks) {
applog(LOG_WARNING, "GPU #%d: only %u mem clocks available (p-states)", dev_id, nclocks);
}
for (uint8_t u=0; u < nclocks; u++) { for (uint8_t u=0; u < nclocks; u++) {
// ordered by pstate (so highest P0 first) // ordered by pstate (so highest P0 first)
if (u == wanted_pstate) { if (u == wanted_pstate) {
mem_clk = clocks[u]; mem_clk = mem_clocks[u];
break; break;
} }
} }
uint32_t* gpu_clocks = NULL;
nclocks = 0; nclocks = 0;
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, NULL); nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, NULL);
nclocks = min(nclocks, 127); if (nclocks) {
if (nclocks) gpu_clocks = (uint32_t*) calloc(1, sizeof(uint32_t) * nclocks + 4);
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, clocks); rc = nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, gpu_clocks);
for (uint8_t u=0; u < nclocks; u++) { if (rc == NVML_SUCCESS) {
// ordered desc, so get first // ordered desc, get the max app clock (do not limit)
if (clocks[u] <= gpu_clk) { gpu_clk = gpu_clocks[0];
gpu_clk = clocks[u];
break;
} }
free(gpu_clocks);
} }
rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[n], mem_clk, gpu_clk); rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[n], mem_clk, gpu_clk);
if (rc != NVML_SUCCESS) { if (rc != NVML_SUCCESS) {
applog(LOG_WARNING, "GPU #%d: pstate %s", dev_id, nvmlh->nvmlErrorString(rc)); applog(LOG_WARNING, "GPU #%d: pstate P%d (%u/%u) %s", dev_id, (int) wanted_pstate,
mem_clk, gpu_clk, nvmlh->nvmlErrorString(rc));
return -1; return -1;
} }