gpuminer0 workLoop 1 #3db89d40 #3db89d40 main Mining on PoWhash #b2b1b835 : 24641536 H/s = 12320768 hashes / 0.5 s main Mining on PoWhash #b2b1b835 : 25115592 H/s = 25165824 hashes / 1.002 s main Mining on PoWhash #b2b1b835 : 25256398 H/s = 38010880 hashes / 1.505 s main Mining on PoWhash #b2b1b835 : 25339280 H/s = 50855936 hashes / 2.007 s main Mining on PoWhash #b2b1b835 : 25274441 H/s = 63438848 hashes / 2.51 s main Mining on PoWhash #b2b1b835 : 25326661 H/s = 76283904 hashes / 3.012 s main Mining on PoWhash #b2b1b835 : 25356745 H/s = 89128960 hashes / 3.515 s
Not sure how mining with the quadro should work so I am not sure if I need to run multiple instances of eth -G for it to work with all available cores. Using ANY--opencl-device reports the same card.
# eth -M -G Benchmarking on platform: { "platform": "NVIDIA CUDA", "device": "Quadro K4200", "version": "OpenCL 1.1 CUDA" } Preparing DAG... ℹ 16:44:56|gpuminer0 workLoop 0 #00000000… #00000000… ℹ 16:44:56|gpuminer0 Initialising miner... Using platform: NVIDIA CUDA Warming up... Using device: Quadro K4200(OpenCL 1.1 CUDA) Trial 1... 5854549 Trial 2... 5854549 Trial 3... 5854549 Trial 4... 5767168 Trial 5... 5854549 min/mean/max: 5767168/5837072/5854549 H/s inner mean: 1951516 H/s Phoning home to find world ranking... Error phoning home. ET is sad.
These numbers were achieved running @Genoil's ethcudaminer executable during CUDA performance testing. Command to follow similar procedure: ./ethcudaminer.exe -U -M
Hey, I've just released a version of my cuda miner that also has some benefits for all you AMD / OpenCL folks. Check out the description (and Windows binaries) here. Unfortunately it doesn't have dag chunk stuff in there yet.
The values listed for the R9 280 have now changed. (I'm guessing due to a slight mining algorithm modification.) The value I am seeing now on my card is ~20MH/s, that is a 5MH/s jump from what it was two months ago and easily makes the R9 280 shine. Typical R9 280s go for around 230, whereas the R9 290 goes for $300+ and only operates around 24MH/s.
Comments
$ ethminer -G -M
Benchmarking on platform: 2-thread CPU
Preparing DAG...
Warming up...
Trial 1... 66
Trial 2... 33
Trial 3... 66
Trial 4... 66
Trial 5... 66
min/mean/max: 33/59/66 H/s
inner mean: 22 H/s
Then if I try adding --opencl-devices (tried up to 12), it does the same.
Is there a way list OpenCL device IDs?
gpuminer0 workLoop 1 #3db89d40 #3db89d40
main Mining on PoWhash #b2b1b835 : 24641536 H/s = 12320768 hashes / 0.5 s
main Mining on PoWhash #b2b1b835 : 25115592 H/s = 25165824 hashes / 1.002 s
main Mining on PoWhash #b2b1b835 : 25256398 H/s = 38010880 hashes / 1.505 s
main Mining on PoWhash #b2b1b835 : 25339280 H/s = 50855936 hashes / 2.007 s
main Mining on PoWhash #b2b1b835 : 25274441 H/s = 63438848 hashes / 2.51 s
main Mining on PoWhash #b2b1b835 : 25326661 H/s = 76283904 hashes / 3.012 s
main Mining on PoWhash #b2b1b835 : 25356745 H/s = 89128960 hashes / 3.515 s
min/mean/max: 0/3.2/16.3 MH/s
inner mean: 5.4 MH/s
Actual average while mining: 15.6MH/s
102.5 MH/s
This gold is hard to mine.
min/mean/max: .349525/12.705245/15.816021 MH/s
inner mean: 10.514887 MH/s
min/mean/max: 0 / 12932437 / 17388885 H/s (very first try)
inner mean: 15786894 H/s (about the actual rate)
min/mean/max: 5767168/5837072/5854549 H/s
inner mean: 1951516 H/s
Not sure how mining with the quadro should work so I am not sure if I need to run multiple instances of
eth -G
for it to work with all available cores. Using ANY--opencl-device
reports the same card.Benchmarking on platform: { "platform": "AMD Accelerated Parallel Processing", "device": "Hawaii", "version": "OpenCL 1.2 AMD-APP (1642.5)" }
Preparing DAG...
i 15:31:14|gWpaurmmiinnge ru0p . .w.o
rkLoop 0 #00000000… #00000000…
i 15:31:14|gpuminer0 Initialising miner...
Found 1 platforms
Using platform: AMD Accelerated Parallel Processing
Using device: Hawaii(OpenCL 1.2 AMD-APP (1642.5))
Trial 1... 25768903
Trial 2... 25760319
Trial 3... 25760319
Trial 4... 25672996
Trial 5... 25672996
min/mean/max: 25672996/25727106/25768903 H/s
inner mean: 17176407 H/s
Phoning home to find world ranking...
Error phoning home. ET is sad.
min/mean/max: 24117248/24169676/24292010 H/s
inner mean: 16165546 H/s
min/mean/max: 0/10317994/13893632 H/s
inner mean: 12,565,446 H/s
min/mean/max: 4893354/11822694/12670293 H/s
inner mean: 3156650 H/s
Everyone else in this thread seems to get ~25 MH/s for 280X's, while mine only does exactly half of that.
Any ideas? Anyone else experienced this?
i7 ~4 GHz
Samsung 850 Evo Pro 500 GB SSD (3D-NAND Version)
GTX 980
RAM 8 GB
min/mean/max: 18,087,936/18,315,127/1,852,482 H/s
inner mean: 18,408,333 H/s
These numbers were achieved running @Genoil's ethcudaminer executable during CUDA performance testing. Command to follow similar procedure: ./ethcudaminer.exe -U -M
Benchmarking on platform: { "platform": "CUDA 7.0", "device": "GeForce GTX 780", "version": "Compute 3.5" }
min/mean/max: 18000554/18122066/18270906 H/s
inner mean: 18168958 H/s
min/mean/max: 0/32.715571/42.205184 MH
inner mean: 40.457557 MH
Trial 3... 42117802
Trial 4... 42205184
Trial 5... 42205184
min / mean / max: 70997 H/s / 102391 H/s / 111381 H/s
Inner mean: 109120 H/s
Does this seem correct?
Benchmarking on platform: { "platform": "CUDA 7.0", "device": "GeForce GTX 750", "version": "Compute 5.0" }
Preparing DAG...
Warming up...
ℹ 11:43:37|cudaminer0 workLoop 0 #00000000… #00000000…
ℹ 11:43:37|cudaminer0 Initialising miner...
Using device: GeForce GTX 750(5.0)
Trial 1... 8388608
Trial 2... 8388608
Trial 3... 8475989
Trial 4... 8388608
Trial 5... 8388608
min/mean/max: 8388608/8406084/8475989 H/s
GPU 0: AMD Gigabyte WindForce R9-290 (slightly overclocked)
Stats:
min/mean/max: 524288/20967241/26214400 H/s
inner mean: 26032506 H/s
tldr; 26.032506 MH/s
min / mean / max: 23.592960/23.662864/23.767722 MH/s
Inner mean: 15.816021 MH/s
Hi, everyone! I'm new here.. just managed to get this done.
Is it ok?? How to make it better??