could someone please help. this is my command for normal scrypt.
u xxx.xxx -p xxx --worksize 256 --lookup-gap 2 --thread-concurrency 8192 -g 2 --intensity 12 -Q 0
i have a vtx3d 7870 le tahiti and i get 345khash/s.
could some please tell me how to change it for vertcoin. thanks.
Topic: The big VTC Vertcoin settings thread - page 21. (Read 264260 times)
For anyone who cant get over 8192 thread concurrency this is what finally fixed it for me.
Source: gaywallets post here:
https://litecointalk.org/index.php?topic=6842.105
What he said:
For anyone who encounters this same issue with setx gpu_max_alloc_percent 100 not working...
You can check the registry edit in HKEY_CURRENT_USER\ENVIRONMENT... if the DWORD value there is not in ALL CAPS, you must delete it. Then re-open command prompt and send the command again, but with ALL CAPS. It would appear that the AMD APP SDK does not recognize it unless it is in all caps, even though the registry won't overwrite if you re-pass with all caps. You must delete the old key first. Fun stuff.
I'll play with threadcount and get back to this thread later with my heat results. Quick googling shows for about $50 you can pick up a GELID ICY REV 2 and drop temps even when max OC down to 70-80, even in a case. I'll have to test the stability at high temps with a single gpu thread and a high threadcount for stability before aftermarket modifications (even if I have to run the fan at 90 or something ridiculously loud)... with multiple threads it seems to get a lot of HW NONCE errors (which from what I gather is basically shared memory getting overwritten by the other cores, causing invalid submissions).
Source: gaywallets post here:
https://litecointalk.org/index.php?topic=6842.105
What he said:
For anyone who encounters this same issue with setx gpu_max_alloc_percent 100 not working...
You can check the registry edit in HKEY_CURRENT_USER\ENVIRONMENT... if the DWORD value there is not in ALL CAPS, you must delete it. Then re-open command prompt and send the command again, but with ALL CAPS. It would appear that the AMD APP SDK does not recognize it unless it is in all caps, even though the registry won't overwrite if you re-pass with all caps. You must delete the old key first. Fun stuff.
I'll play with threadcount and get back to this thread later with my heat results. Quick googling shows for about $50 you can pick up a GELID ICY REV 2 and drop temps even when max OC down to 70-80, even in a case. I'll have to test the stability at high temps with a single gpu thread and a high threadcount for stability before aftermarket modifications (even if I have to run the fan at 90 or something ridiculously loud)... with multiple threads it seems to get a lot of HW NONCE errors (which from what I gather is basically shared memory getting overwritten by the other cores, causing invalid submissions).
cant set both shaders and threads. if you set the tc it ignores the shaders. if you set only the shaders without the tc if will auto calculate what the tc should be.
--thread-concurrency:
This tunes the optimal size of work that scrypt can do. It is internally tuned
by cgminer to be the highest reasonable multiple of shaders that it can
allocate on your GPU. Ideally it should be a multiple of your shader count.
vliw5 architecture (R5XXX) would be best at 5x shaders, while VLIW4 (R6xxx and
R7xxx) are best at 4x. Setting thread concurrency overrides anything you put
into --shaders and is ultimately a BETTER way to tune performance.
SUMMARY: Spend lots of time finding the highest value that your device likes
and increases hashrate.
--thread-concurrency:
This tunes the optimal size of work that scrypt can do. It is internally tuned
by cgminer to be the highest reasonable multiple of shaders that it can
allocate on your GPU. Ideally it should be a multiple of your shader count.
vliw5 architecture (R5XXX) would be best at 5x shaders, while VLIW4 (R6xxx and
R7xxx) are best at 4x. Setting thread concurrency overrides anything you put
into --shaders and is ultimately a BETTER way to tune performance.
SUMMARY: Spend lots of time finding the highest value that your device likes
and increases hashrate.
It's an ambiguous misphrasing. When you set --shaders cgminer will automatically derive a thread concurrency value from that number but you can specify your shader count and following that override the auto-derived TC by setting your own value. If I set 256 shaders I get the same performance as setting 256 shaders and 1024 TC but a different performance if I set 256 shader and 512 TC, and if I only set TC 1024 I get a different performance as well, and different still if I set 128 shaders and/or 1024/512 TC.
Yeah I tried that, doesn't do much good. Do you think I'd do better if I switch from windows 7 64bit to ubuntu ?
I don't think soI have the same card, this is my setting/result :
setx GPU_MAX_ALLOC_PERCENT 100 (in the .bat)
you can see 0 HW
It can always be a hardward issue, some card are just not good
Yeah I tried that, doesn't do much good. Do you think I'd do better if I switch from windows 7 64bit to ubuntu ?
With 7950 i'm getting now about 260khash, but I get accepted shares: 75900 ; rejected 64, hw errors 331
Is this ratio normal? For some reason i'm getting hw errors at intensity 13, if i lower to intensity 12, errors stop but the hash rate falls by 100khash
Is this ratio normal? For some reason i'm getting hw errors at intensity 13, if i lower to intensity 12, errors stop but the hash rate falls by 100khash
you may try to lower the thread concurrency
With 7950 i'm getting now about 260khash, but I get accepted shares: 75900 ; rejected 64, hw errors 331
Is this ratio normal? For some reason i'm getting hw errors at intensity 13, if i lower to intensity 12, errors stop but the hash rate falls by 100khash
Is this ratio normal? For some reason i'm getting hw errors at intensity 13, if i lower to intensity 12, errors stop but the hash rate falls by 100khash
I get about 443 KH running with the Sapphire R9 290X reference card. Just have these extra options in my bat file
-I 20 -g 1
-I 20 -g 1
Any Reject or HW error @ -I 20 43 KH?
Guys, i need your help! i have got 3 machines and none of them can run intesnity above 12! i get tons of HW.
At first i thought that it was a RAM problem, then i have added 4 more gigs for a total of 8 gigs, but it didnt help!
I have 280x,7950 and r9 290. None of them can run intensity above 12 with out HW.
I have tried adding setx commands etc.
Lowering Core, memory. Does not help.
Im using vertminer 0.5.3 also tried 0.5.2
Any ideas?
i would really love to mine this coin, but with intensity 12, i get around 200khs on each card
At first i thought that it was a RAM problem, then i have added 4 more gigs for a total of 8 gigs, but it didnt help!
I have 280x,7950 and r9 290. None of them can run intensity above 12 with out HW.
I have tried adding setx commands etc.
Lowering Core, memory. Does not help.
Im using vertminer 0.5.3 also tried 0.5.2
Any ideas?
i would really love to mine this coin, but with intensity 12, i get around 200khs on each card
TC less than 8192?
whats an igpu, is that integrated?
i can confirm lowering the core indeed give some 20 to 30 odd KHZ per card (for me), and adding the --shaders xxxx option gives a few more!
https://lh6.googleusercontent.com/-Qol4_sjSK1s/UvzOaA5mabI/AAAAAAAAGmE/Z1FEBv28qzE/w715-h434-no/vert2.jpg
TC: 8192
shaders: 2048
I 13
G 2
Core/Mem: 1030/1500
again plz igonre the GPU 1 (iGPU)
https://lh6.googleusercontent.com/-Qol4_sjSK1s/UvzOaA5mabI/AAAAAAAAGmE/Z1FEBv28qzE/w715-h434-no/vert2.jpg
TC: 8192
shaders: 2048
I 13
G 2
Core/Mem: 1030/1500
again plz igonre the GPU 1 (iGPU)
Guys, i need your help! i have got 3 machines and none of them can run intesnity above 12! i get tons of HW.
At first i thought that it was a RAM problem, then i have added 4 more gigs for a total of 8 gigs, but it didnt help!
I have 280x,7950 and r9 290. None of them can run intensity above 12 with out HW.
I have tried adding setx commands etc.
Lowering Core, memory. Does not help.
Im using vertminer 0.5.3 also tried 0.5.2
Any ideas?
i would really love to mine this coin, but with intensity 12, i get around 200khs on each card
At first i thought that it was a RAM problem, then i have added 4 more gigs for a total of 8 gigs, but it didnt help!
I have 280x,7950 and r9 290. None of them can run intensity above 12 with out HW.
I have tried adding setx commands etc.
Lowering Core, memory. Does not help.
Im using vertminer 0.5.3 also tried 0.5.2
Any ideas?
i would really love to mine this coin, but with intensity 12, i get around 200khs on each card
hello dear, i need help
using Binaries for Linux: http://sourceforge.net/projects/cpuminer/files/pooler-cpuminer-2.3.2-linux-x86_64.tar.gz (x86-64)
os: Linux new 2.6.18-348.4.1.el5.centos.plus #1 SMP Tue Apr 16 18:51:49 EDT 2013 x86_64 x86_64 x86_64 GNU/Linux
./minerd --url=stratum+tcp://s.eu.vertcoin.org:3333 --userpass=Plus.s741:coolmegapass
and work problem:
Failed to get Stratum session id
and
DEBUG: reject reason: Share is above target
printscreen http://clip2net.com/s/6Lfb2Z
booooo/booooo/booooo/booooo/booooo ((((
using Binaries for Linux: http://sourceforge.net/projects/cpuminer/files/pooler-cpuminer-2.3.2-linux-x86_64.tar.gz (x86-64)
os: Linux new 2.6.18-348.4.1.el5.centos.plus #1 SMP Tue Apr 16 18:51:49 EDT 2013 x86_64 x86_64 x86_64 GNU/Linux
./minerd --url=stratum+tcp://s.eu.vertcoin.org:3333 --userpass=Plus.s741:coolmegapass
and work problem:
Failed to get Stratum session id
and
DEBUG: reject reason: Share is above target
printscreen http://clip2net.com/s/6Lfb2Z
booooo/booooo/booooo/booooo/booooo ((((
LOWERING THE CORE
The core engine frequency that gives highest hashrate will, however, go up proportionate to your memory clock. And what's really interesting, the achievable ratio between the two seems to be dependent on how low a thread concurrency you can get before WU drops off or hardware errors begin. There are places in this state space where it will seem like increasing TC gives you a little extra hashrate, but really you want it low so that you can get effective hashing with the engine clock as high as possible.
Or at least that's what I seem to have observed with my 290's. I'm still not entirely sure - the numbers move in mysterious ways, sometimes.
I mined away happily for 24 hours yesterday on my 6x R9 280x rig with TC@8192, core@1030, mem@1500, -g2, lookup gap 2... giving 370Kh/s per card. 5 HW Errors after 24 hours was acceptable. But this morning I woke up to a DEAD card, and I couldn't even restart the miner, it would just hang.
It's quite a shame, I really like VTC but I'm working full time during the day with no access to my mining box (company policy... can't use remote desktop etc)... and I need the thing to be reliable.
I got 7850 and R9 270
7850 is running at 180Khash
270 is running at 202
end of the day after reset stats etc. 7850 is doing alot better share count then 270. 7850 I hardly see any stale or rejects.
I'm having a lot of trouble running 2x CGminer at once.. One will start and another will stop or crash the card.
I let the card idle for 5+min and both will restart fine, sometime.
Its weird how settings are different from one user to another
7850 is running at 180Khash
270 is running at 202
end of the day after reset stats etc. 7850 is doing alot better share count then 270. 7850 I hardly see any stale or rejects.
I'm having a lot of trouble running 2x CGminer at once.. One will start and another will stop or crash the card.
I let the card idle for 5+min and both will restart fine, sometime.
Its weird how settings are different from one user to another
LOWERING THE CORE
The core engine frequency that gives highest hashrate will, however, go up proportionate to your memory clock. And what's really interesting, the achievable ratio between the two seems to be dependent on how low a thread concurrency you can get before WU drops off or hardware errors begin. There are places in this state space where it will seem like increasing TC gives you a little extra hashrate, but really you want it low so that you can get effective hashing with the engine clock as high as possible.
Or at least that's what I seem to have observed with my 290's. I'm still not entirely sure - the numbers move in mysterious ways, sometimes.
try lowering the core
i discovered that the only value that matter is the core, leave everything as it was with the previous settings
Just to make sure everyone saw this. It's VERY important.i discovered that the only value that matter is the core, leave everything as it was with the previous settings
On normal scrypt, my r9 280x's sweetspot is 1085/1500.
I was getting less than 300 Kh/s in vertminer with these values.
LOWERING THE CORE TO 1030 MADE ME GAIN OVER 70Kh/s.
I'm now happily mining away at 370 per card with a total of 4GB of ram...
7950 has 1792 shader units, so try setting --shaders 1792 and experiment with a thread concurrency that is a multiple of that, i.e. 3584, 7168, 8960 etc. Also try updating to a newer Catalyst driver set. I got better results on Catalyst 13.9 than the old 13.4 or something I had first.
--shaders 1792 --thread-concurrency 7168 --lookup-gap 0 --I 13 --g 2 --w 286
cant set both shaders and threads. if you set the tc it ignores the shaders. if you set only the shaders without the tc if will auto calculate what the tc should be.
--thread-concurrency:
This tunes the optimal size of work that scrypt can do. It is internally tuned
by cgminer to be the highest reasonable multiple of shaders that it can
allocate on your GPU. Ideally it should be a multiple of your shader count.
vliw5 architecture (R5XXX) would be best at 5x shaders, while VLIW4 (R6xxx and
R7xxx) are best at 4x. Setting thread concurrency overrides anything you put
into --shaders and is ultimately a BETTER way to tune performance.
SUMMARY: Spend lots of time finding the highest value that your device likes
and increases hashrate.
Huh, barely got the hang of cgminer. Is it similar at least? Alternatively which card should I buy that will support two gpu threads( budget versions), please don't recommend r9 290 or 280
7950 has 1792 shader units, so try setting --shaders 1792 and experiment with a thread concurrency that is a multiple of that, i.e. 3584, 7168, 8960 etc. Also try updating to a newer Catalyst driver set. I got better results on Catalyst 13.9 than the old 13.4 or something I had first.
Thanks man, got +5kh with catalyst 13.9. Not much but hey everything counts
My settings on sapphire dualx 7950 are:
gpu 1023, memory 1225, 1,1v set with sapphire trixx
--shaders 1792 --thread-concurrency 7168 --lookup-gap 0 --I 13 --g 2 --w 286
getting 265 kh
One question, if i add r9 270, can i run two different settings (for different cards) in vertminer?
Always when i got gpu 0 failure it was a problem with too high of thread concurrency. If you set too high gpu threads it freezes the system, same with intensity
you can always use --g 1,2 --engine-core 1000,900 etc to tweak the settings for each individual card.
You cant run two cards on different number of threads (-g) on cgminer. Only Sgminer allows that.
Jump to: