08-09-2018, 02:58 PM
I've did some test with hashcat v4.2.1 with an AMD RX480 GPU on Windows 7 with the latest AMD "Adrenalin" driver ( win7-64bit-radeon-pro-software-adrenalin-edition-18.7.1-july9.exe ) just to make sure it's working and we do not need to change anything to the driver recommendation on the main hashcat page.
Some guy on IRC mentioned to change it which is the reason I wanted to verify what he said.
I haven't used AMD hardware on a Windows box (and much more important, the AMD original driver) for quite some time, but I can confirm everything still works out of the box.
Here's a screenshot from afterburner with hashcat running in benchmark. Each time a new benchmark (for a different hash mode) is starting the graph reset, the GPU utilization ideally goes to 100% and the core clock stays on maximum.
Here's some robust benchmark output:
The most interessting fact is that the windows driver is clearly faster than the linux driver (MD5 @ Linux: 10,203,053,705 vs Windows: 11,981,034,475). I think this is related to the power limit that I can set to 50% on windows using afterburner. Important! Do not confuse setting the power limit with overclocking. Setting the power limit just effectively disables the power consumption throttling. This enables the benchmark to run smooth on the 1266 Mhz boost clock and not to jump around crazy to save a few watts. The 1266 Mhz is the stock clock setting for this GPU. Damn AMD, why can't you just give us such a quality driver for Linux!
- atom
Some guy on IRC mentioned to change it which is the reason I wanted to verify what he said.
I haven't used AMD hardware on a Windows box (and much more important, the AMD original driver) for quite some time, but I can confirm everything still works out of the box.
Quote:Platform ID #2
Vendor : Advanced Micro Devices, Inc.
Name : AMD Accelerated Parallel Processing
Version : OpenCL 2.1 AMD-APP (2671.3)
Device ID #2
Type : GPU
Vendor ID : 1
Vendor : Advanced Micro Devices, Inc.
Name : Ellesmere
Version : OpenCL 2.0 AMD-APP (2671.3)
Processor(s) : 36
Clock : 1266
Memory : 3264/4096 MB allocatable
OpenCL Version : OpenCL C 2.0
Driver Version : 2671.3
Here's a screenshot from afterburner with hashcat running in benchmark. Each time a new benchmark (for a different hash mode) is starting the graph reset, the GPU utilization ideally goes to 100% and the core clock stays on maximum.
Here's some robust benchmark output:
Quote:d:\tools\hashcat-4.2.1>hashcat64.exe -b --mac
# version: v4.2.1-3-gdad05d9f
# option: --optimized-kernel-enable
2:0:1266:1750:49.92:11981034475
2:100:1266:1750:70.20:4258718563
2:1400:1266:1750:81.78:1838697478
2:1700:1266:1750:82.91:449761650
2:2500:1264:1750:90.30:191428
2:1000:1266:1750:51.76:23059660320
2:3000:1266:1750:51.21:11564387348
2:5500:1266:1750:91.23:13241766084
2:5600:1264:1750:89.10:856523284
2:1500:1266:1750:82.41:452750315
2:500:1266:1750:58.43:4795802
2:3200:1266:1750:62.15:8988
2:1800:1266:1750:96.96:75429
2:7500:1266:1750:76.88:121269534
2:13100:1266:1750:78.08:119613469
2:15300:1266:1750:89.86:34430
2:15900:1266:1750:66.19:17582
2:7100:1266:1750:65.94:2963
2:11600:1266:1750:50.14:5196
2:12500:1244:1750:46.97:14522
2:13000:1266:1750:49.64:16366
2:6211:1266:1750:58.91:145405
2:13400:1266:1750:146.84:170447
2:6800:1265:1750:43.07:1441705
2:11300:1266:1750:91.07:2046
Started: Thu Aug 09 14:51:20 2018
Stopped: Thu Aug 09 14:54:59 2018
The most interessting fact is that the windows driver is clearly faster than the linux driver (MD5 @ Linux: 10,203,053,705 vs Windows: 11,981,034,475). I think this is related to the power limit that I can set to 50% on windows using afterburner. Important! Do not confuse setting the power limit with overclocking. Setting the power limit just effectively disables the power consumption throttling. This enables the benchmark to run smooth on the 1266 Mhz boost clock and not to jump around crazy to save a few watts. The 1266 Mhz is the stock clock setting for this GPU. Damn AMD, why can't you just give us such a quality driver for Linux!
- atom