GPU miners comparsion

[pafnucy]

Tested on Tesla K20c / Xeon E5 2620 / 16GB RAM / Win7 x64:

[miner] [version/options] [cpm]
girino OpenCL jhProtominer gpuv6 480
girino OpenCL jhProtominer gpuv6 -m2048 580
girino OpenCL jhProtominer gpuv4 630
girino OpenCL jhProtominer gpuv4 -m2048 640
girino OpenCL jhProtominer gpuv2 410

cudaptswin-SM35 870

cudaptswin-0.4-exp 1250->1160

arCUDAminer 1.0c avx 1180
arCUDAminer 1.0c cm3 avx 1250->1180
arCUDAminer 1.0c highperf avx 1050
arCUDAminer 1.0c highperf cm3 avx around 1070

PtsGPUz 0.3c lit up my antivirus so disqualified

[dga]

2200 cpm on my gtx titan with my SM35 varient of DGAS code with a improved sha512 core.

three thumbs up.

A good place to admit that I didn't think this optimization would be worth doing (and my naive test of it didn't work), and I was completely wrong.  go go open source version.  *grin*

[slothlike]

2200 cpm on my gtx titan with my SM35 varient of DGAS code with a improved sha512 core.



https://dl.dropboxusercontent.com/u/33838/cudaptswin-0.2-SM35.7z

[daem0n]

GTX 570

PtsGPUz0.3b - 1030 CPM
arCUDAminer1.0c - 780 CPM

[paul]

arCUDAminer1.0c is the best  :D
Thank you archit

GTX690 = 2150 c/m


GTS450 = 253 c/m


Quadro 2000 = 204 c/m

[archit]

arCUDAMiner v1.0c GTX 560 ti - 540 cpm
arCUDAMiner v1.0c GTX 650 ti - 720 cpm

[earthbound]

Please update the spreadsheet which I link to from this post:

https://bitsharestalk.org/index.php?topic=2437.0

--with anything I'm missing. I'm not sure whether I'd put any speed tests in it; maybe that could be another spreadsheet?

[COOLERbyPSP]

1gh v1.2 - windows 8.1 64 bit - 6950@70 2gb (940/1320) ~1180 cpm
                                             - 7970 3gb (1125/1575) ~1020 cpm
1gh v1.2 - windows 7 64 bit - 6950 1gb (850/1200) ~80 cpm (FFFFFFFFUUUUUUUU)

[drin]

And the PC is still usable, except for watching movies.

Use Quick Sync ;)

[emski]

CPU: I7 3770K
RAM: 16GB DDR3 1600mhz
GTX 570 (no OC)
Windows 8 64

PtsGPUz0.2ab.exe - collisions/min: 830.5174 Shares total: 4301 *

* note these results are while the following CPU miner is running:
Software: yam-yvg1900-M7i-win64-ivy-bridge
Threads:5
PTS Agg. CPM: 269.0;(1024 MB)(AV=9 after AV-optimization)

And the PC is still usable, except for watching movies.

[dga]

What is "cudapts"?  shrug

I am getting 1750 cpm with gtx 780 +150 core -100 memory on "PtsGPUz0.2ab"

ptsgpuz0.2ab is a windows port that combines jhprotominer with the GPU core code from cudapts.  Your performance should be similar to that achieved by cudapts, at least as of a version or two ago.

[zvs]

What is "cudapts"?  shrug

I am getting 1750 cpm with gtx 780 +150 core -100 memory on "PtsGPUz0.2ab"

[reorder]

1gh v1.2 - windows 8 64 bit.  sapphire 7970 ghz edition 3gb, 1040 core 1500 ram clock speed. ~1120 CPM

Just for fun:

cudapts (linux) v2014-01-12 GTX690, stock clock settings, 1780 c/m.

Do you mind to try 1gh miner on this one? That would give a general idea how much OpenCL is inferior to CUDA.

Unfortunately, that machine also has one of my coin wallets on it, and I don't run untrusted binaries on it. :(

(Mind you, I'm not saying there's anything wrong with the 1gh miner - but there's been a lot of malware floating around.)

But - probably the best case thus far is the EC2 results showing 480 c/m vs about 780.  One might expect the same per-core results to extend to the 690, so maybe it would get about 1150?  (total SWAG)

I totally understand your concern about running unknown binaries, yet, maybe there is a way to arrange it securely without much hassle? For example, in VM with mapped PCI device if you are already using such setup for development?

480cpm must be something else, 1gh miner yields 687cpm on g2.2xlarge and I believe there is no more room for optimization (but it is possible that different Nvidia OpenCL compiler versions may produce different results).

[dga]

1gh v1.2 - windows 8 64 bit.  sapphire 7970 ghz edition 3gb, 1040 core 1500 ram clock speed. ~1120 CPM

Just for fun:

cudapts (linux) v2014-01-12 GTX690, stock clock settings, 1780 c/m.

Do you mind to try 1gh miner on this one? That would give a general idea how much OpenCL is inferior to CUDA.

Unfortunately, that machine also has one of my coin wallets on it, and I don't run untrusted binaries on it. :(

(Mind you, I'm not saying there's anything wrong with the 1gh miner - but there's been a lot of malware floating around.)

But - probably the best case thus far is the EC2 results showing 480 c/m vs about 780.  One might expect the same per-core results to extend to the 690, so maybe it would get about 1150?  (total SWAG)

[dga]

1gh v1.2 - windows 8 64 bit.  sapphire 7970 ghz edition 3gb, 1040 core 1500 ram clock speed. ~1120 CPM

Just for fun:

cudapts (linux) v2014-01-12 GTX690, stock clock settings, 1780 c/m.

Do you mind to try 1gh miner on this one? That would give a general idea how much OpenCL is inferior to CUDA.

Unfortunately, that machine also has one of my coin wallets on it, and I don't run untrusted binaries on it. :(

(Mind you, I'm not saying there's anything wrong with the 1gh miner - but there's been a lot of malware floating around.)