MrFreeDragon
|
|
August 26, 2019, 01:57:11 PM |
|
i got RX 580 110-120 Mkey/sec with no boost
Thanks. That's reasonable. So the cards RX 580/570/480 are 2 times less effecient rather than RTX 2080ti. Five cards RX 580/570/480 will cost as one card RTX 2080ti, however have the total key scan rate 2 times less - approx. 600 mKey/sec Something strange with GTX 1080ti - it cost just 1.5-2 times less than RTX 2080ti, however produces only 150Mkey/sec, but should at least 500Mkey/sec, i guess. Here is that guy: https://bitcointalk.org/index.php?topic=5166284.msg52246277#msg52246277Probably, he had very wrong settings...
|
|
|
|
vimp666
Jr. Member
Offline
Activity: 37
Merit: 1
|
|
August 26, 2019, 02:09:50 PM Last edit: August 26, 2019, 02:44:17 PM by vimp666 |
|
I read again most of the results posted here, and collected them to one table. Of corse results could differ from card to card depending on the settings. I also add the price of the CUDA. I tried to collect some benchmark. If you have some other device - pls, do not hesitate to add 1. RTX 2080ti 1,200-1,300M key/sec (price $1,200-1,500) 2. GTX 1070 250 Mkey/sec (price $450-550) 3. GTX 1080ti 150 Mkey/sec (price $700-1,000) 4. GTX 680 109 Mkey/sec (price $150) 5. RX 480 107 Mkey/sec (price $130-150) 6. RX 470 105 Mkey/sec (price $150-300) 7. RX 580 89 Mkey/sec (price $200-250) 8. RX 560 50 Mkey/sec (price $100-150) 9. R9 280/290x 20 Mkey/sec (price $50)
i got RX 580 110-120 Mkey/sec with no boost GTX 1080ti 150 Mkey/sec ?probably just someone did not configure it.I think this is wrong, it should also produce about 900-1000Mkey/s on compressed addresses.Productivity 2080ti percent is 30 higher .
|
|
|
|
bulleteyedk
Jr. Member
Offline
Activity: 95
Merit: 3
|
|
August 26, 2019, 02:30:46 PM Last edit: August 26, 2019, 03:03:38 PM by bulleteyedk |
|
I have been running Bitcrack on Windows so far, but im planning to use it on Ubuntu 18.04, im just having trouble installing either the Zielar/brichard19 version make is running fine for BitCrack make BUILD_CUDA=1 gives me a fatal error terminating that install? how to solve this? In file included from cudaUtil.cpp:1:0:0 cudaUtil.h:4:10: fatal error: cuda.h: No such file or folder #include <cuda.h> ^-------- compilation terminated.
CUDA is installed together with geforce drivers
|
|
|
|
paschok81
Newbie
Offline
Activity: 14
Merit: 0
|
|
August 26, 2019, 07:47:42 PM |
|
I am at 1080ti maximum reached 380 Mkey / s. This is probably normal relative to 1070 and AMD. I wonder why 2080ti is more powerful in ~ 3 times than 1080ti, it should be about ~ 30% (500 Mkey / s). Due to what such a gap, have thoughts?
|
|
|
|
MrFreeDragon
|
|
August 26, 2019, 08:32:35 PM |
|
I am at 1080ti maximum reached 380 Mkey / s. This is probably normal relative to 1070 and AMD. I wonder why 2080ti is more powerful in ~ 3 times than 1080ti, it should be about ~ 30% (500 Mkey / s). Due to what such a gap, have thoughts?
RTX 2080ti has 4,352 data flow processors, GTX 1080ti has 3,584 data flow processors. So, it is reasonable that GTX 1080ti should have at least 800-900 Mkey/sec. If we have a look through "market price" terms, so the minimum should be 500-600 MKey/sec. Can you try to change and test different parameters b, t, p for block, threads and points? The number b (blocks) try to put 2-3 times more than the number of computing items of your card. But anyway, thank you for your answer. The previous one I had was just 150 MKey/sec with GTX 1080ti whch seems totally insufficient and wrong.
|
|
|
|
zielar (OP)
|
|
August 26, 2019, 11:10:00 PM |
|
I have been running Bitcrack on Windows so far, but im planning to use it on Ubuntu 18.04, im just having trouble installing either the Zielar/brichard19 version make is running fine for BitCrack make BUILD_CUDA=1 gives me a fatal error terminating that install? how to solve this? In file included from cudaUtil.cpp:1:0:0 cudaUtil.h:4:10: fatal error: cuda.h: No such file or folder #include <cuda.h> ^-------- compilation terminated.
CUDA is installed together with geforce drivers In your console in Ubuntu type this: export PATH=/usr/local/cuda-10.1/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-10.1/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
|
If you want - you can send me a donation to my BTC wallet address 31hgbukdkehcuxcedchkdbsrygegyefbvd
|
|
|
bulleteyedk
Jr. Member
Offline
Activity: 95
Merit: 3
|
|
August 27, 2019, 12:05:12 AM |
|
Thanks for much needed help with the build problem, it did'nt solve the issue, but i think i have made a mess trying too much, with different versions of BitCrack I think it def. have something to do with some wrong library paths, like you mentioned...
|
|
|
|
Raj Lassi
Newbie
Offline
Activity: 17
Merit: 1
|
|
August 27, 2019, 01:54:29 AM |
|
IMHO ranges should go from blah...blah...00000000 to blah...blah...FFFFFFFF just to keep things neat and tidy. maybe i am OCD.
why are all these bizarre ranges being scanned? like 2b87eefd01e43a40 - 2b87f0798ea43a40 it makes no sense to me.
|
|
|
|
Raj Lassi
Newbie
Offline
Activity: 17
Merit: 1
|
|
August 27, 2019, 02:02:31 AM |
|
IMHO ranges should go from blah...blah...00000000 to blah...blah...FFFFFFFF just to keep things neat and tidy. maybe i am OCD.
why are all these bizarre ranges being scanned? like 2b87eefd01e43a40 - 2b87f0798ea43a40 it makes no sense to me.
is something being converted from decimal? that makes even less sense. i just don't get it. enlighten me.
|
|
|
|
MeBender
Jr. Member
Offline
Activity: 114
Merit: 5
|
|
August 27, 2019, 03:52:32 AM |
|
IMHO ranges should go from blah...blah...00000000 to blah...blah...FFFFFFFF just to keep things neat and tidy. maybe i am OCD.
why are all these bizarre ranges being scanned? like 2b87eefd01e43a40 - 2b87f0798ea43a40 it makes no sense to me.
is something being converted from decimal? that makes even less sense. i just don't get it. enlighten me. Yeah so that range is within the 62-bit range, what is the problem?
|
CryptoCoin - The latest Cuckoo Cycle coin - https://crypt-o-coin.cash Github: https://github.com/GonzoTheDev
|
|
|
MrFreeDragon
|
|
August 27, 2019, 11:58:58 AM |
|
I tried to test clbittrack on old laptop, and found some interesting. It works on integrated laptop graphics card with speed 12-14MKey/sec, but on integrated CPU with speed just 0.8-1MKey/sec. HOWEVER, CPU made real calclulations and found the solution (for #61 wallet), but graphics card showed 15times more speed, but found nothing. I selected very short range with the know key --> pls have a look at the screen shot. How could it be? Why clbitcrack with the graphic card just wasting the time? So, the quiestion is also - could that speed showed with RTX2080ti devices (1,200-1,300 MKey/sec) also be just waste of time without real search?
|
|
|
|
paniker
Newbie
Offline
Activity: 49
Merit: 0
|
|
August 27, 2019, 01:16:59 PM |
|
I tried to test clbittrack on old laptop, and found some interesting. It works on integrated laptop graphics card with speed 12-14MKey/sec, but on integrated CPU with speed just 0.8-1MKey/sec. HOWEVER, CPU made real calclulations and found the solution (for #61 wallet), but graphics card showed 15times more speed, but found nothing. I selected very short range with the know key --> pls have a look at the screen shot. https://i.ibb.co/64V7nXH/61test.jpgHow could it be? Why clbitcrack with the graphic card just wasting the time? So, the quiestion is also - could that speed showed with RTX2080ti devices (1,200-1,300 MKey/sec) also be just waste of time without real search? i tested at almost founded addreses, and it find, but only on 1. hm maybe you make some mistakes in commands? is it real?
|
|
|
|
GoldTiger69
|
|
August 27, 2019, 01:26:49 PM |
|
I tried to test clbittrack on old laptop, and found some interesting. It works on integrated laptop graphics card with speed 12-14MKey/sec, but on integrated CPU with speed just 0.8-1MKey/sec. HOWEVER, CPU made real calclulations and found the solution (for #61 wallet), but graphics card showed 15times more speed, but found nothing. I selected very short range with the know key --> pls have a look at the screen shot. How could it be? Why clbitcrack with the graphic card just wasting the time? So, the quiestion is also - could that speed showed with RTX2080ti devices (1,200-1,300 MKey/sec) also be just waste of time without real search? If the range is too short, then clbitcrack can miss it. Try again with a larger range, it should work.
|
|
|
|
MrFreeDragon
|
|
August 27, 2019, 01:31:32 PM |
|
I tried to test clbittrack on old laptop, and found some interesting. It works on integrated laptop graphics card with speed 12-14MKey/sec, but on integrated CPU with speed just 0.8-1MKey/sec. HOWEVER, CPU made real calclulations and found the solution (for #61 wallet), but graphics card showed 15times more speed, but found nothing. I selected very short range with the know key --> pls have a look at the screen shot. How could it be? Why clbitcrack with the graphic card just wasting the time? So, the quiestion is also - could that speed showed with RTX2080ti devices (1,200-1,300 MKey/sec) also be just waste of time without real search? If the range is too short, then clbitcrack can miss it. Try again with a larger range, it should work. Have a look at the screen shot - the range was the same as for CPU, so for graphics card. CPU found the key, but the graphicas card - not. So i worry that the search with graphics card could be waste of time on such situation. PS. The higher range (268 million combinations) the same issue: graphics card did not see the key with the speed 5Mkey, but CPU found with the speed 0.8MKey/sec.
|
|
|
|
vimp666
Jr. Member
Offline
Activity: 37
Merit: 1
|
|
August 27, 2019, 01:38:14 PM Last edit: August 27, 2019, 01:48:31 PM by vimp666 |
|
I tried to test clbittrack on old laptop, and found some interesting. It works on integrated laptop graphics card with speed 12-14MKey/sec, but on integrated CPU with speed just 0.8-1MKey/sec. HOWEVER, CPU made real calclulations and found the solution (for #61 wallet), but graphics card showed 15times more speed, but found nothing. I selected very short range with the know key --> pls have a look at the screen shot. How could it be? Why clbitcrack with the graphic card just wasting the time? So, the quiestion is also - could that speed showed with RTX2080ti devices (1,200-1,300 MKey/sec) also be just waste of time without real search? if you meant clBitCrack, then in the old versions there were a lot of bugs(he was skip keys) on the Intel and AMD windows graphics. but in release 0.26 it seems to be fixed, and in release 0.30 there was an additional fix for Intel graphics.all this is connected with the implementation of OCL manufacturer. possibly on some devices errors remained. https://github.com/brichard19/BitCrack/issues/81
|
|
|
|
zielar (OP)
|
|
August 27, 2019, 03:30:56 PM |
|
I tried to test clbittrack on old laptop, and found some interesting. It works on integrated laptop graphics card with speed 12-14MKey/sec, but on integrated CPU with speed just 0.8-1MKey/sec. HOWEVER, CPU made real calclulations and found the solution (for #61 wallet), but graphics card showed 15times more speed, but found nothing. I selected very short range with the know key --> pls have a look at the screen shot. How could it be? Why clbitcrack with the graphic card just wasting the time? So, the quiestion is also - could that speed showed with RTX2080ti devices (1,200-1,300 MKey/sec) also be just waste of time without real search? i tested at almost founded addreses, and it find, but only on 1. hm maybe you make some mistakes in commands? is it real? I cannot confirm any issues... This is my test from few moments ago:
|
If you want - you can send me a donation to my BTC wallet address 31hgbukdkehcuxcedchkdbsrygegyefbvd
|
|
|
MrFreeDragon
|
|
August 27, 2019, 04:00:51 PM |
|
I cannot confirm any issues... This is my test from few moments ago: Thanks... You tested with Intel - it works. I tested with AMD A8 with graphics AMD Radeon R5. And used the latest release v.30 With graphics skips the key and shows no result
|
|
|
|
bulleteyedk
Jr. Member
Offline
Activity: 95
Merit: 3
|
|
August 27, 2019, 08:50:38 PM |
|
This #62 range is killing us I thought it would be cool to visualise the scanned ranges (at least the ones people have reported) The layout is 1900x1000, each block shown is 3 pixels, and represents a decimal range of 35900556513413. A white block is not scanned and a black has a range that have been scanned (not necessary the hole decimal range) starting block is 7CCE5EFDACCF6807 and range is shown up to one number before #63: 7CCE5EFDACCF6807 You might want to zoom a bit into this image, it's pretty hard to actually see those small scanned areas
|
|
|
|
zielar (OP)
|
|
August 27, 2019, 09:00:48 PM |
|
This #62 range is killing us I thought it would be cool to visualise the scanned ranges (at least the ones people have reported) The layout is 1900x1000, each block shown is 3 pixels, and represents a decimal range of 35900556513413. A white block is not scanned and a black has a range that have been scanned (not necessary the hole decimal range) starting block is 7CCE5EFDACCF6807 and range is shown up to one number before #63: 7CCE5EFDACCF6807 You might want to zoom a bit into this image, it's pretty hard to actually see those small scanned areas First of all - the range is not really small ... Three months pass as I started scanning, and with the strength of 1590Mkey / s multiplied by over 100 processes - I still did not give the correct key. Nevertheless, you flew over the band with this range ... where did you get it? Scope in DEC: 1497845136656582919 - 8993229949524469767 - according to you 2305843009213693952 - 4611686018427387903 - according to reality
|
If you want - you can send me a donation to my BTC wallet address 31hgbukdkehcuxcedchkdbsrygegyefbvd
|
|
|
MrFreeDragon
|
|
August 27, 2019, 09:06:28 PM |
|
Imagine that there was a mistake and #62 wallet has not 62bit key...
|
|
|
|
|