Discrepancy between Benchmark numbers and actual numbers
#1
I recently stood up a 4 GPU system, and noticed something strange. When I run a benchmark -

hashcat64.exe -b -m1000

Quote:OpenCL Platform #1: NVIDIA Corporation
======================================
* Device #1: GeForce RTX 2080, 2048/8192 MB allocatable, 46MCU
* Device #2: GeForce GTX 1080 Ti, 2816/11264 MB allocatable, 28MCU
* Device #3: GeForce GTX 1070, 2048/8192 MB allocatable, 15MCU
* Device #4: GeForce GTX 980 Ti, 1536/6144 MB allocatable, 22MCU

OpenCL Platform #2: Intel(R) Corporation
========================================
* Device #5: Intel(R) HD Graphics 4600, skipped.
* Device #6: Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz, skipped.

Benchmark relevant options:
===========================
* --optimized-kernel-enable

Hashmode: 1000 - NTLM

Speed.#1.........: 69485.6 MH/s (10.40ms) @ Accel:16 Loops:1024 Thr:1024 Vec:2
Speed.#2.........: 39607.4 MH/s (23.18ms) @ Accel:32 Loops:1024 Thr:1024 Vec:2
Speed.#3.........: 36190.7 MH/s (27.03ms) @ Accel:64 Loops:1024 Thr:1024 Vec:2
Speed.#4.........: 31403.6 MH/s (22.95ms) @ Accel:32 Loops:1024 Thr:1024 Vec:2
Speed.#*.........:  176.7 GH/s

Started: Tue Aug 27 14:41:16 2019
Stopped: Tue Aug 27 14:41:31 2019


I get those numbers.
But when actually cracking NTLM hashes, I'll see something very different.

hashcat64.exe -m1000 -a3 -w4

Quote:Session..........: hashcat
Status...........: Quit
Hash.Type........: NTLM
Hash.Target......: (redacted)
Time.Started.....: Tue Aug 27 14:48:11 2019 (10 secs)
Time.Estimated...: Tue Aug 27 14:49:49 2019 (1 min, 28 secs)
Guess.Mask.......: ?1?1?1?1?1?1?1?1?1?1?1 [11]
Guess.Charset....: (redacted)
Guess.Queue......: 1/1 (100.00%)
Speed.#1.........: 14942.9 MH/s (4.96ms) @ Accel:16 Loops:128 Thr:1024 Vec:1
Speed.#2.........:  9676.2 MH/s (10.58ms) @ Accel:32 Loops:128 Thr:1024 Vec:1
Speed.#3.........:  7701.1 MH/s (7.40ms) @ Accel:64 Loops:64 Thr:1024 Vec:1
Speed.#4.........:  8809.2 MH/s (9.49ms) @ Accel:32 Loops:128 Thr:1024 Vec:1
Speed.#*.........: 41129.3 MH/s
Recovered........: 1/5 (20.00%) Digests, 0/1 (0.00%) Salts
Progress.........: 412132311040/4049565169664 (10.18%)
Rejected.........: 0/412132311040 (0.00%)
Restore.Point....: 144474112/1475789056 (9.79%)
Restore.Sub.#1...: Salt:0 Amplifier:1152-1280 Iteration:0-128
Restore.Sub.#2...: Salt:0 Amplifier:1408-1536 Iteration:0-128
Restore.Sub.#3...: Salt:0 Amplifier:1664-1728 Iteration:0-64
Restore.Sub.#4...: Salt:0 Amplifier:1664-1792 Iteration:0-128
Candidates.#1....: kDKkwk!wkdd -> w!LdLw3rr!0
Candidates.#2....: DDrkkWrdWKL -> Rw!r0w!wkdd
Candidates.#3....: !Ll0wk$k$wk -> w$03$RRKkrd
Candidates.#4....: !LlRD0$wWKL -> $lDrkWrdWKL
Hardware.Mon.#1..: Temp: 38c Fan: 71% Util: 92% Core:2010MHz Mem:6800MHz Bus:1
Hardware.Mon.#2..: Temp: 28c Fan: 40% Util: 97% Core:1075MHz Mem:5005MHz Bus:1
Hardware.Mon.#3..: Temp: 43c Fan: 97% Util: 95% Core:2037MHz Mem:3802MHz Bus:1
Hardware.Mon.#4..: Temp: 44c Fan:100% Util: 97% Core:1379MHz Mem:3304MHz Bus:1


I'm aware that multiple hashes in the list can slow it down, but I get the same numbers with 1 or 5 (which is in the list in this case). I can re-run with 1 hash if needed, but does anyone have any ideas as to why I'm seeing SUCH a huge difference? I could understand a few % points slower, but not THAT much of a difference.

I realize there's quite possibly something I'm missing here, but I've been chasing this for the last few days and I'm out of ideas.
Reply


Messages In This Thread
Discrepancy between Benchmark numbers and actual numbers - by XakEp - 08-27-2019, 10:52 PM