Gigabyte GTX 960 OC 2GB OR Sapphire R9 270X 2G?
#1
There is a sale on around my neck of the woods and these are quite similar prices. I've seen that the GTX 960 is quite similar to the 270x in the speeds on http://golubev.com/gpuest.htm although the SHA1 speed is faster and MD5 is slower. My question is, what would be better for hash cracking as a card from the lower end of the high range cards.
#2
That page is a good ressource to start with, although under oclhashcat everything can shift again.
You can roughly calculate the hash speed by comparing the number of shaders and multiply the values.

GTX960
(Speed value of) GTX980 * 0,5
GTX 970 * 0,61
GTX 750Ti * 1,6

270x
290x * 0,45
280x/7970 * 0,625

Since performance and price of both cards is similar I'd go for the GTX960. Way less power and heat, great overclocking. Remember the benchmarks here are mostly run under stock speed, a plus 20% in overclocking means 20% higher values. If you stick to AMD I'd more recommend the R9 285 over the 270x because of less power/heat but 40% more shaders. Or get a used 7970 aka 280x.
#3
Thanks for that, went out and brought a GTX 960. I'm having an odd issue with fan speed showing as "N/A" and the fans not going past standard rates, but I'll probably figure that out.
#4
Great if you could post some benchmarks!
#5
Sure!

cudaHashcat v1.33 starting in benchmark-mode...

Device #1: GeForce GTX 960, 2048MB, 1278Mhz, 8MCU

Hashtype: MD4
Workload: 1024 loops, 256 accel

Speed.GPU.#1.: 12297.6 MH/s

Hashtype: MD5
Workload: 1024 loops, 256 accel

Speed.GPU.#1.: 6012.6 MH/s

Hashtype: SHA1
Workload: 1024 loops, 256 accel

Speed.GPU.#1.: 1883.2 MH/s

Hashtype: SHA256
Workload: 1024 loops, 256 accel

Speed.GPU.#1.: 755.6 MH/s

Hashtype: SHA384
Workload: 256 loops, 256 accel

Speed.GPU.#1.: 241.5 MH/s

Hashtype: SHA512
Workload: 256 loops, 256 accel

Speed.GPU.#1.: 234.1 MH/s

Hashtype: SHA-3(Keccak)
Workload: 128 loops, 32 accel

Speed.GPU.#1.: 181.8 MH/s

Hashtype: RipeMD160
Workload: 1024 loops, 256 accel

Speed.GPU.#1.: 1277.9 MH/s

Although it does end up crashing with "ERROR: cuStreamSynchronize() 999" which I think I've fixed the issue but will reboot and upload the rest of the bench tomorrow.