Bitcoin Forum
August 19, 2017, 10:40:33 PM *
News: Latest stable version of Bitcoin Core: 0.14.2  [Torrent].
 
   Home   Help Search Donate Login Register  
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 [24] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 ... 1143 »
  Print  
Author Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX]  (Read 3220772 times)
peacefulmind
Full Member
***
Offline Offline

Activity: 196


View Profile
April 22, 2013, 08:43:39 PM
 #461

I am using 4/22 version.  Same cmd as before, however total has dropped from 520 to 410kH.

Same clocks here is .bat

cudaminer.exe --url http://127.0.0.1:8332/ --userpass xxx.x:123 -i 0,0 -d 0,1 -m 1,1 -C 2,2 -l 84x4,84x4


This got 520kH on the 2x titan in 4/17 release.

Quote from: FrictionlessCoin
"I think you are to hung up on this notion about 'pre-mining' being a No-No."
- from journeys into the dark depths of the alt coin forum....
1503182433
Hero Member
*
Offline Offline

Posts: 1503182433

View Profile Personal Message (Offline)

Ignore
1503182433
Reply with quote  #2

1503182433
Report to moderator
Advertised sites are not endorsed by the Bitcoin Forum. They may be unsafe, untrustworthy, or illegal in your jurisdiction. Advertise here.
1503182433
Hero Member
*
Offline Offline

Posts: 1503182433

View Profile Personal Message (Offline)

Ignore
1503182433
Reply with quote  #2

1503182433
Report to moderator
1503182433
Hero Member
*
Offline Offline

Posts: 1503182433

View Profile Personal Message (Offline)

Ignore
1503182433
Reply with quote  #2

1503182433
Report to moderator
tacotime
Legendary
*
Offline Offline

Activity: 1484



View Profile
April 22, 2013, 08:45:28 PM
 #462

Do you feel you are at release candidate level yet?  I want to add this to guiminer-scrypt when it hits maturity.

Code:
XMR: 44GBHzv6ZyQdJkjqZje6KLZ3xSyN1hBSFAnLP6EAqJtCRVzMzZmeXTC2AHKDS9aEDTRKmo6a6o9r9j86pYfhCWDkKjbtcns
Misiolap
Newbie
*
Offline Offline

Activity: 14


View Profile
April 22, 2013, 08:45:53 PM
 #463

Isn't ulong the same as uint on 32 bit builds?
On 64bit linux it breaks things, because ulong is 64bit and uint is 32bit.

My tip jars: 1FQMFpqnCH1ATPGoHLnmSAXB1gBh9vAKXC | LZiXcRvUr5wJrkfvc9vXvqDwe51YfVPRkv
cbuchner1
Hero Member
*****
Offline Offline

Activity: 742


View Profile
April 22, 2013, 08:49:05 PM
 #464

Do you feel you are at release candidate level yet?  I want to add this to guiminer-scrypt when it hits maturity.

hmm I am probably not going to change the console and command line options output now. But stability (error checking) has to be improved before this can even hit beta status.
cbuchner1
Hero Member
*****
Offline Offline

Activity: 742


View Profile
April 22, 2013, 08:50:11 PM
 #465

Isn't ulong the same as uint on 32 bit builds?
On 64bit linux it breaks things, because ulong is 64bit and uint is 32bit.

Any suggestion for a portable 32 bit type among 32 bit and 64 bit builds?  I thought int changed size depending on architecture, long is always 32 bits, and long long is always 64 bits.

EDIT: I've been reading up on the differences between Microsoft's LLP64 model vs. Unix/Linux LP64 model. I will have to change a few things in the code, then.

Christian
cbuchner1
Hero Member
*****
Offline Offline

Activity: 742


View Profile
April 22, 2013, 08:51:29 PM
 #466

I am using 4/22 version.  Same cmd as before, however total has dropped from 520 to 410kH.
This got 520kH on the 2x titan in 4/17 release.

Wah! I need to empty a bottle of wine now.
K1773R
Legendary
*
Offline Offline

Activity: 1610


/dev/null


View Profile
April 22, 2013, 08:57:42 PM
 #467

wait, 1 titan is 7kh/s slower as my 580? that's sad Sad

[GPG Public Key]  [Devcoin Builds]  [BBQCoin Builds]  [Multichain Blockexplorer]  [Multichain Blockexplorer - PoS Coins]  [Ufasoft Miner Linux Builds]
BTC/DVC/TRC/FRC: 1K1773RbXRZVRQSSXe9N6N2MUFERvrdu6y ANC/XPM AK1773RTmRKtvbKBCrUu95UQg5iegrqyeA NMC: NK1773Rzv8b4ugmCgX789PbjewA9fL9Dy1 LTC: LKi773RBuPepQH8E6Zb1ponoCvgbU7hHmd EMC: EK1773RxUes1HX1YAGMZ1xVYBBRUCqfDoF BQC: bK1773R1APJz4yTgRkmdKQhjhiMyQpJgfN
Misiolap
Newbie
*
Offline Offline

Activity: 14


View Profile
April 22, 2013, 09:18:41 PM
 #468

Any suggestion for a portable 32 bit type among 32 bit and 64 bit builds?  I thought int changed size depending on architecture, long is always 32 bits, and long long is always 64 bits.

EDIT: I've been reading up on the differences between Microsoft's LLP64 model vs. Unix/Linux LP64 model. I will have to change a few things in the code, then.

Christian

For general purpose vars use uint*_t from <stdint.h>

I'm not sure what should be used for CUDA vector types for portability.

On 64bit linux:
sizeof(ulong2): 16, sizeof(uint2): 8

On 32bit linux it's probably 8 for both.

My tip jars: 1FQMFpqnCH1ATPGoHLnmSAXB1gBh9vAKXC | LZiXcRvUr5wJrkfvc9vXvqDwe51YfVPRkv
peacefulmind
Full Member
***
Offline Offline

Activity: 196


View Profile
April 22, 2013, 09:18:48 PM
 #469

I am using 4/22 version.  Same cmd as before, however total has dropped from 520 to 410kH.
This got 520kH on the 2x titan in 4/17 release.

Wah! I need to empty a bottle of wine now.


Christian,

Perhaps it is something in my setup?  This is not made for a mining rig, I use it for day to day and gaming.

990x
12GB RAM
2x Titan
5760x1200 SLI
64bit win7

I have noticed when I set interactive to 1,1 it freezes, also when I try to let it auto-tune it freezes.

I have a ton of games and applications on this machine - so it may be my system.

Perhaps the new 4/22 build needs different settings than the ones I used on the 4/17 build?  I will try some more.

My dedicated mining machines are lean and mean and using AMD RADEON on Linux/a few win7 - so it is hard to compare.

You are trailblazing new ground!

Quote from: FrictionlessCoin
"I think you are to hung up on this notion about 'pre-mining' being a No-No."
- from journeys into the dark depths of the alt coin forum....
cbuchner1
Hero Member
*****
Offline Offline

Activity: 742


View Profile
April 22, 2013, 09:28:12 PM
 #470

Okay, this would be the 2nd attempt for the day.

uint32_t becomes typedef'd as unsigned int
ulong2 becomes uint2
ulong4 becomes uint4

and Titan kernel now does uint2 based memory transactions in a shared memory buffer of [16+2] width, which should reduce warp serialization.

now where's my bottle of wine?
Misiolap
Newbie
*
Offline Offline

Activity: 14


View Profile
April 22, 2013, 10:04:43 PM
 #471

Additionally shared buffers for 64bit builds must be 64bit aligned.

If it's worth to save some memory for 32bit builds something like this can be done:

Code:
#if __x86_64__
#define _64BIT_ALIGN 1
#else
#define _64BIT_ALIGN 0
#endif

And for each buffer:

Code:
__shared__ uint32_t X[WARPS_PER_BLOCK][WU_PER_WARP][32+1+_64BIT_ALIGN];

My tip jars: 1FQMFpqnCH1ATPGoHLnmSAXB1gBh9vAKXC | LZiXcRvUr5wJrkfvc9vXvqDwe51YfVPRkv
cbuchner1
Hero Member
*****
Offline Offline

Activity: 742


View Profile
April 22, 2013, 10:22:09 PM
 #472

Additionally shared buffers for 64bit builds must be 64bit aligned.

Does this also apply when targeting compute_10, sm_10 (which is done in salsa_kernel.cu) ?

Christian
Misiolap
Newbie
*
Offline Offline

Activity: 14


View Profile
April 22, 2013, 10:48:47 PM
 #473

Yes, without this on 64bit it kernel dies due to Warp Misaligned Address on compute_10.

My tip jars: 1FQMFpqnCH1ATPGoHLnmSAXB1gBh9vAKXC | LZiXcRvUr5wJrkfvc9vXvqDwe51YfVPRkv
cbuchner1
Hero Member
*****
Offline Offline

Activity: 742


View Profile
April 22, 2013, 10:50:58 PM
 #474

Yes, without this on 64bit it kernel dies due to Warp Misaligned Address on compute_10.

ok then I'll put that in and make this a final upload for today. Thank you for the explanations and for developing the patch.
Misiolap
Newbie
*
Offline Offline

Activity: 14


View Profile
April 22, 2013, 11:18:01 PM
 #475

Great, now it works out-of-box for me (salsa_kernel), thanks.

My tip jars: 1FQMFpqnCH1ATPGoHLnmSAXB1gBh9vAKXC | LZiXcRvUr5wJrkfvc9vXvqDwe51YfVPRkv
dbabo
Jr. Member
*
Offline Offline

Activity: 41


View Profile
April 22, 2013, 11:29:09 PM
 #476

Yes, without this on 64bit it kernel dies due to Warp Misaligned Address on compute_10.

ok then I'll put that in and make this a final upload for today. Thank you for the explanations and for developing the patch.

well deserved bottle of wine shall be corked:
Code:
[2013-04-22 19:29:03] 1 miner threads started, using 'scrypt' algorithm.
[2013-04-22 19:29:14] GPU #0: GeForce GT 430 with compute capability 2.1
[2013-04-22 19:29:14] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
[2013-04-22 19:29:14] GPU #0: Performing auto-tuning (Patience...)
[2013-04-22 19:29:21] GPU #0:   24.34 khash/s with configuration  4x6
[2013-04-22 19:29:21] GPU #0: using launch configuration  4x6
[2013-04-22 19:29:21] GPU #0: GeForce GT 430, 4608 hashes, 0.26 khash/s
[2013-04-22 19:29:21] GPU #0: GeForce GT 430, 1536 hashes, 15.62 khash/s
[2013-04-22 19:29:25] GPU #0: GeForce GT 430, 78336 hashes, 22.29 khash/s
[2013-04-22 19:29:30] GPU #0: GeForce GT 430, 112128 hashes, 22.11 khash/s
[2013-04-22 19:29:35] GPU #0: GeForce GT 430, 110592 hashes, 21.79 khash/s
[2013-04-22 19:29:40] GPU #0: GeForce GT 430, 109056 hashes, 21.37 khash/s
[2013-04-22 19:29:45] GPU #0: GeForce GT 430, 107520 hashes, 22.23 khash/s

and no i686 deps Smiley
first coin goes to you
thank you!
nst6563
Sr. Member
****
Offline Offline

Activity: 252


View Profile
April 22, 2013, 11:48:33 PM
 #477

Yes, without this on 64bit it kernel dies due to Warp Misaligned Address on compute_10.

ok then I'll put that in and make this a final upload for today. Thank you for the explanations and for developing the patch.

well deserved bottle of wine shall be corked:
Code:
[2013-04-22 19:29:03] 1 miner threads started, using 'scrypt' algorithm.
[2013-04-22 19:29:14] GPU #0: GeForce GT 430 with compute capability 2.1
[2013-04-22 19:29:14] GPU #0: interactive: 1, tex-cache: 0 , single-alloc: 0
[2013-04-22 19:29:14] GPU #0: Performing auto-tuning (Patience...)
[2013-04-22 19:29:21] GPU #0:   24.34 khash/s with configuration  4x6
[2013-04-22 19:29:21] GPU #0: using launch configuration  4x6
[2013-04-22 19:29:21] GPU #0: GeForce GT 430, 4608 hashes, 0.26 khash/s
[2013-04-22 19:29:21] GPU #0: GeForce GT 430, 1536 hashes, 15.62 khash/s
[2013-04-22 19:29:25] GPU #0: GeForce GT 430, 78336 hashes, 22.29 khash/s
[2013-04-22 19:29:30] GPU #0: GeForce GT 430, 112128 hashes, 22.11 khash/s
[2013-04-22 19:29:35] GPU #0: GeForce GT 430, 110592 hashes, 21.79 khash/s
[2013-04-22 19:29:40] GPU #0: GeForce GT 430, 109056 hashes, 21.37 khash/s
[2013-04-22 19:29:45] GPU #0: GeForce GT 430, 107520 hashes, 22.23 khash/s

and no i686 deps Smiley
first coin goes to you
thank you!


If you overclock that gt430 a bit you can get in the 30kh/s range. 
I currently get 36kh/s on my gt430 with configuration 20x8. 

Was I helpful? BTC: 1B5SzPsBDuGLvGYuWw5b4Po8MaFSGck1Dt    LTC: LSQEM9re967ds5ZTHxs6SUvmKVpB5KdMwZ    DOGE:  DS3HN4s4YainUsiwrFcAoXa4e3WiSaM74P
dbabo
Jr. Member
*
Offline Offline

Activity: 41


View Profile
April 22, 2013, 11:53:23 PM
 #478

...
If you overclock that gt430 a bit you can get in the 30kh/s range. 
I currently get 36kh/s on my gt430 with configuration 20x8. 


and how do i do that?  Roll Eyes
nst6563
Sr. Member
****
Offline Offline

Activity: 252


View Profile
April 23, 2013, 12:05:07 AM
 #479

...
If you overclock that gt430 a bit you can get in the 30kh/s range. 
I currently get 36kh/s on my gt430 with configuration 20x8. 


and how do i do that?  Roll Eyes

Google a tool called NvidiaInspector (I think it's from TechPowerup).  It will let you adjust the fan speeds, voltage, and clock speeds of the core/mem/shader of most all nvidia cards.

I use it to set the clocks on my gt430 card to 882Mhz core, 810Mhz mem, 1760Mhz shader, .990v.  That combo yields between 36kh/s-38kh/s.  If I go any higher than that I get the driver crash.

Your mileage may vary though on the clock speeds you can attain.  I have an EVGA GT430 so I'm not sure how it compares to other flavors.

Was I helpful? BTC: 1B5SzPsBDuGLvGYuWw5b4Po8MaFSGck1Dt    LTC: LSQEM9re967ds5ZTHxs6SUvmKVpB5KdMwZ    DOGE:  DS3HN4s4YainUsiwrFcAoXa4e3WiSaM74P
dbabo
Jr. Member
*
Offline Offline

Activity: 41


View Profile
April 23, 2013, 12:38:30 AM
 #480


Google a tool called NvidiaInspector (I think it's from TechPowerup).  It will let you adjust the fan speeds, voltage, and clock speeds of the core/mem/shader of most all nvidia cards.
..
that seems to be win only. i don't have win pcs.
but there seemed to be a nvclock utility that should have some options. I'll look into it. Thank you for pointing out to the opportunity to OC.
Pages: « 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 [24] 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 ... 1143 »
  Print  
 
Jump to:  

Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!