Encryption Projects as SU group

ViaHussun

Donating Member
Messages
4,099
Something is seriously wrong then, to be getting such terrible speeds
You see my results, on a nearly 10 year old I7 with a gtx 980

Which GPU driver are you using?
It's a laptop.
TDP speed limit 75 watts
Maybe that's why it's slow.
 

Attachments

  • Ekran görüntüsü 2023-08-29 201724.png
    Ekran görüntüsü 2023-08-29 201724.png
    41.6 KB · Views: 20

sdrfgs

Registered
Messages
116
Oh yes that explains it, yes laptops have weak gpus due to power requirements.

If you can disable power limits for the gpu to improve your speeds Try that.

a desktop with 3060 gets 1060 million keys at stock with 1 instance running ...3 times the speed of yours
 

ViaHussun

Donating Member
Messages
4,099
Oh yes that explains it, yes laptops have weak gpus due to power requirements.

If you can disable power limits for the gpu to improve your speeds Try that.

a desktop with 3060 gets 1060 million keys at stock with 1 instance running ...3 times the speed of yours
I will still try to support
 

sattechtips

Senior Member
Messages
429
i have 2 archives

all .exe 1267 KB
Cudabiss22464.exe
Cudabiss24064.exe
Cudabiss25664.exe
Cudabiss44864.exe
Cudabiss48064.exe
Cudabiss51264.exe

cudart.dll
cudart32_32_16.dll
Cudart64_31_9.dll
cutil32.dll
cutil64.dll
input.txt
packets_TS.exe (19kb)
----------------
Then this which called itself Cudabiss 2.1.2

all .exe 2126 KB
Cudabiss96064.exe
Cudabiss89664.exe
Cudabiss51264.exe
Cudabiss48064.exe
Cudabiss44864.exe
Cudabiss12864.exe
Cudabiss12064.exe
Cudabiss11264.exe
cudart.dll
cutil64.dll
cudart64_40_17.dll

It usually crashes for me with this error

c:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.0/C/src/template/template.cu(5062) : cudaSafeCall() Runtime API error : unspecified launch failure.
Cudabiss96064.exe
Cudabiss89664.exe
cudart.dll
cutil64.dll
cudart64_40_17.dll
Upload this to see what speed will give, must be most fast from all versions.
 

sdrfgs

Registered
Messages
116
Cudabiss96064.exe
Cudabiss89664.exe
cudart.dll
cutil64.dll
cudart64_40_17.dll
Upload this to see what speed will give, must be most fast from all versions.
When im done testing it, before this use to crash on my 980, but a couple of months ago i upgraded this pc from win 7 to win 10 and hadn't tried it again.
I forgot that I had this version several years
 

sdrfgs

Registered
Messages
116
It looks to me that cudabiss 51264 based on cudart64_31_9.dll has been compiled with
c:\ProgramData\NVIDIA Corporation\NVIDIA GPU Computing SDK\C\src\simpleAtomicIntrinsics\template_kernel.cu

Then some dudes, obviously with access to cudabiss source files, managed to go even further and added support to newer versions of cudart64_65_14.dll and cudart64_75_18.dll .
The new ones have been compiled with
C:/ProgramData/NVIDIA Corporation/CUDA Samples/v7.0/0_Simple/TemplateCB/template.cu

Well that version should be the best of all? but who has it?
 

dvlajkovic

Well Known Member
Messages
502
We are trying to make something better, not to run the ones that still take many hours to cover full range.
 

sdrfgs

Registered
Messages
116
Well actually i agree, but we need to establish a baseline of how the current cudabiss performs and its limits and weaknesses , before dreaming about building something better.

The fact that you need to run 3 or 4 copys to max out the performance shows, cudabiss could have at minimum a 4x speed boost with probably minimul code changes for modern hardware
 

sdrfgs

Registered
Messages
116
there is a lot of assembly code inside the cudabiss exe files if you open them in notepad
öŠì…SEÞ»~
š*Â^Z25œ¨s0PíUº ‰ À Êy 8 ` Y E @ -dlcm=cg -maxrregcount=63 c:/ProgramData/NVIDIA Corporation/NVIDIA GPU Computing SDK 4.0/C/src/template/template.cu .version 2.3
.target sm_20
.address_size 64
// compiled with C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v4.0\\bin/../open64/lib//be.exe
// nvopencc 4.0 built on 2011-05-13

.visible .func _Z15stream_cypher11iPhS_S_iPiS0_S0_h (.param .s32 __cudaparmf1__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u64 __cudaparmf2__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u64 __cudaparmf3__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u64 __cudaparmf4__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .s32 __cudaparmf5__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u64 __cudaparmf6__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u64 __cudaparmf7__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u64 __cudaparmf8__Z15stream_cypher11iPhS_S_iPiS0_S0_h, .param .u32 __cudaparmf9__Z15stream_cypher11iPhS_S_iPiS0_S0_h)

.visible .func _Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_ (.param .s32 __cudaparmf1__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf2__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf3__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf4__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .s32 __cudaparmf5__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf6__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf7__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf8__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u32 __cudaparmf9__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf10__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf11__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf12__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf13__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf14__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf15__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf16__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_, .param .u64 __cudaparmf17__Z16stream_cypher111iPhS_S_iPiS0_S0_hS_S_S_S_S_S_S_S_)

.visible .func _Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_ (.param .s32 __cudaparmf1__Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_, .param .u64 __cudaparmf2__Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_, .param .u64 __cudaparmf3__Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_, .param .u64 __cudaparmf4__Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_, .param .u64 __cudaparmf5__Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_, .param .u64 __cudaparmf6__Z22stream_cypher_bs32_cpuiPhPiS_S0_S0_)
 

sdrfgs

Registered
Messages
116
$LDWbegin__Z15stream_cypher11iPhS_S_iPiS0_S0_h:
ld.param.u32 %r1, [__cudaparmf1__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s32 %r2, %r1;
ld.param.u64 %rd1, [__cudaparmf2__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s64 %rd2, %rd1;
ld.param.u64 %rd3, [__cudaparmf3__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s64 %rd4, %rd3;
ld.param.u64 %rd5, [__cudaparmf4__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s64 %rd6, %rd5;
ld.param.u32 %r3, [__cudaparmf5__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s32 %r4, %r3;
ld.param.u64 %rd7, [__cudaparmf6__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s64 %rd8, %rd7;
ld.param.u64 %rd9, [__cudaparmf7__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s64 %rd10, %rd9;
ld.param.u64 %rd11, [__cudaparmf8__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
mov.s64 %rd12, %rd11;
ld.param.u32 %r5, [__cudaparmf9__Z15stream_cypher11iPhS_S_iPiS0_S0_h];
cvt.u8.u32 %r6, %r5;
mov.s32 %r7, 0;
setp.gt.s32 %p1, %r4, %r7;
mov.u32 %r8, 1;


etc
 

cayoenrique

Senior Member
Messages
489
Note:
Continues Brute-force is a process that requires power, power energy turns into HEAT. HEAT is know to destruct electronic devices.

Continues Brute-Force is done in Desktop PC, in general the bigger the desktop the better. What about laptops. Laptops are build for easy transportation and are expected to be Low Power so that batteries last longer. Laptops are not build to dissipate HEAT. So in general you can use the GPU of the LAPTOP for short periods of time. You need to keep track of ALL Temperature measure devices inside Laptop while doing brute-force. Be aware that it is common for a Laptop to over heat resulting in some bad connections at the GPU. User will then notice that the screen gets dark and nothing gets shown on it. Be careful.
 

sdrfgs

Registered
Messages
116
A waste of time using cuda bruteforce on laptops, unless plugged into power with energy saving disabled. Even then performance will be poor as Laptop GPUS are generally weaker than their desktop versions.

You can get an old desktop for $50 - $100 that will perform better even with older series card like 9/10/16/20 series
 

sdrfgs

Registered
Messages
116
Another option could be just write a plugin for hashcat as all the hardwork has already been done in support of multiple cpu/gpu cracking
 
Top