OK, @ 14,14 temps are OK (68,62), rate is 2.3Gh/s
What should I increase 1st, core clock or mem clock?
Topic: [ANN][BLC] Blakecoin Blake-256 for GPU/FPGA With Merged Mined Pools Stable Net - page 97. (Read 409571 times)
May 22, 2014, 04:11:46 PM
May 22, 2014, 04:08:19 PM
I was able to clone the git repository for the miner but can't get it to build. Anybody with some pointers?
Calhil did a build guide to ubuntu here: https://bitcointalksearch.org/topic/m.6764336
Beautiful! Thanks so much. I'll give it a try when I get back in front of my rig.
May 22, 2014, 04:01:20 PM
OK, we're off to a start anyway - slow but cool (like me!)
That looks better
May 22, 2014, 03:58:18 PM
OK, we're off to a start anyway - slow but cool (like me!)
May 22, 2014, 03:47:44 PM
Cheers Calhil,
I'll give that a whirl & report back.
I'll give that a whirl & report back.
May 22, 2014, 03:24:57 PM
Ok. I thought 7990 would show as only 1 gpu in cgminer.
Could you use this config and set core/memory frequencies as low as possible in an external program?
Could you use this config and set core/memory frequencies as low as possible in an external program?
Code:
{
"pools": [
{
"url": "stratum+tcp://eu3.blakecoin.com:3334",
"user": "",
"pass": ""
}
],
"blake256": true,
"gpu-reorder": true,
"device": "0,1",
"intensity": "8,8",
"gpu-threads": "2",
"gpu-fan": "100",
"temp-cutoff": "95",
"temp-overheat": "85",
"temp-target": "75",
"temp-hysteresis": "3"
}
If this proves to be stable, start increasing intensity for both cards by 1: "9,9" then "10,10", etc. When you find highest intensity you can use, start increasing core frequency. "pools": [
{
"url": "stratum+tcp://eu3.blakecoin.com:3334",
"user": "",
"pass": ""
}
],
"blake256": true,
"gpu-reorder": true,
"device": "0,1",
"intensity": "8,8",
"gpu-threads": "2",
"gpu-fan": "100",
"temp-cutoff": "95",
"temp-overheat": "85",
"temp-target": "75",
"temp-hysteresis": "3"
}
May 22, 2014, 03:21:26 PM
4k rpm probably means your fans are working at 100%
Can you verify those temperatures with another application? Maybe theres a bug with drivers or cgminer.
Just to be sure, is GPU0 and GPU1 a single 7990 or are those 2 separate gpus?
Can you verify those temperatures with another application? Maybe theres a bug with drivers or cgminer.
Just to be sure, is GPU0 and GPU1 a single 7990 or are those 2 separate gpus?
*GROWL* Does nobody read more than a couple of posts back?
YES verified by AfterBurner
YES 7990 is 2 gpus on one card
... and yes, I think it's a bug with the blake version of cgminer, I've had NO such probs on any other algo.
use a more simple conf and control the fans with msi afterburner problem is I think you are trying both from cgminer and from msi afterburner and this is getting issues
This is from one of my rigs has a basic conf and control of the cards is done by Sapphire TRIXX
Code:
{
"pools" : [
{
"url" : "stratum+tcp://la1.blakecoin.com:3334",
"user" : "User.Worker",
"pass" : "WorkerPass"
}
],
"intensity" : "9,9,9",
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"no-restart" : true,
"queue" : "5",
"scan-time" : "40",
"worksize" : "256",
"blake256" : true,
"vectors" : "1",
"no-submit-stale": true,
"kernel-path" : "/"
}
"pools" : [
{
"url" : "stratum+tcp://la1.blakecoin.com:3334",
"user" : "User.Worker",
"pass" : "WorkerPass"
}
],
"intensity" : "9,9,9",
"expiry" : "120",
"failover-only" : true,
"gpu-threads" : "2",
"log" : "5",
"no-restart" : true,
"queue" : "5",
"scan-time" : "40",
"worksize" : "256",
"blake256" : true,
"vectors" : "1",
"no-submit-stale": true,
"kernel-path" : "/"
}
May 22, 2014, 03:03:36 PM
4k rpm probably means your fans are working at 100%
Can you verify those temperatures with another application? Maybe theres a bug with drivers or cgminer.
Just to be sure, is GPU0 and GPU1 a single 7990 or are those 2 separate gpus?
Can you verify those temperatures with another application? Maybe theres a bug with drivers or cgminer.
Just to be sure, is GPU0 and GPU1 a single 7990 or are those 2 separate gpus?
*GROWL* Does nobody read more than a couple of posts back?
YES verified by AfterBurner
YES 7990 is 2 gpus on one card
... and yes, I think it's a bug with the blake version of cgminer, I've had NO such probs on any other algo.
May 22, 2014, 02:22:50 PM
4k rpm probably means your fans are working at 100%
Can you verify those temperatures with another application? Maybe theres a bug with drivers or cgminer.
Just to be sure, is GPU0 and GPU1 a single 7990 or are those 2 separate gpus?
Can you verify those temperatures with another application? Maybe theres a bug with drivers or cgminer.
Just to be sure, is GPU0 and GPU1 a single 7990 or are those 2 separate gpus?
May 22, 2014, 02:06:20 PM
tried?
"gpu-fan" : "100,100",
"gpu-fan" : "100,100",
I have the fans @ 100% on afterburner & have no probs like this on X-11, groestl or myr-gr (there are THREE fans on the 7990 but only 2 are listed on AB) for sure they are all working OK. (there are no sides on the PC, I checked them by eye/hand)
if you scroll the monitoring in afterburner it should show each fan rpm, also a quick visual when its running would be useful
As I said, only 2 fans are listed & have done visual/tactile checks, the fans are fine, the rig is fine, the card is fine. I'm not a coder, I'm a real world engineer so logical thought is a big part of my armoury.
THIS IS DRIVING ME FUCKING CRAZY!
using this minimal config
Code:
"blake256": true,
"gpu-reorder": true,
"device": "0",
"intensity": "14,0",
"gpu-threads": "1",
"temp-cutoff": "95",
"temp-overheat": "85",
"temp-target": "75",
"temp-hysteresis": "3"
"gpu-reorder": true,
"device": "0",
"intensity": "14,0",
"gpu-threads": "1",
"temp-cutoff": "95",
"temp-overheat": "85",
"temp-target": "75",
"temp-hysteresis": "3"
I get this...
But when I change the config to
Code:
"device": "1",
"intensity": "0,14",
"intensity": "0,14",
to drive the other gpu I get this
HOW THE HELL can a switched-off gpu get hotter than one being driven hard?
DOES NOT COMPUTE - DOES NOT COMPUTE - DOES NOT COMPUTE -
May 22, 2014, 12:56:50 PM
tried?
"gpu-fan" : "100,100",
"gpu-fan" : "100,100",
I have the fans @ 100% on afterburner & have no probs like this on X-11, groestl or myr-gr (there are THREE fans on the 7990 but only 2 are listed on AB) for sure they are all working OK. (there are no sides on the PC, I checked them by eye/hand)
if you scroll the monitoring in afterburner it should show each fan rpm, also a quick visual when its running would be useful
May 22, 2014, 12:54:34 PM
tried?
"gpu-fan" : "100,100",
"gpu-fan" : "100,100",
I have the fans @ 100% on afterburner & have no probs like this on X-11, groestl or myr-gr (there are THREE fans on the 7990 but only 2 are listed on AB) for sure they are all working OK. (there are no sides on the PC, I checked them by eye/hand)
May 22, 2014, 12:49:47 PM
tried?
"gpu-fan" : "100,100",
with afterburner click user defined should green up and it should apply that 100% also in the settings make sure to check the Synchronize option
"gpu-fan" : "100,100",
with afterburner click user defined should green up and it should apply that 100% also in the settings make sure to check the Synchronize option
May 22, 2014, 12:47:08 PM
Hey Guys, me again.
I REALLY want to luv this algo but it's driving me bats
No matter how I tweak my .config I get thermal runaway on gpu1 of my HD7990
This is one attempt, with I= 11,3, just before I turned off....
Then later, with gpu1 DISABLED & I=0 I got this before shutting down & going back to myr-gr
With the same rig this is how myr-gr behaves, so it's NOT the rig or card.
Is anyone else using an HD7990 who can post a known-good .config for me to try?
I know gpu1 always runs a bit hotter than gpu0 but on blake it's chuffing ridiculous right now
I REALLY want to luv this algo but it's driving me bats
No matter how I tweak my .config I get thermal runaway on gpu1 of my HD7990
This is one attempt, with I= 11,3, just before I turned off....
Then later, with gpu1 DISABLED & I=0 I got this before shutting down & going back to myr-gr
With the same rig this is how myr-gr behaves, so it's NOT the rig or card.
Is anyone else using an HD7990 who can post a known-good .config for me to try?
I know gpu1 always runs a bit hotter than gpu0 but on blake it's chuffing ridiculous right now
only time I have seen anything like that is when my fan failed on one of my cards, normally the temps on Blake-256 are lower than most algo's
it could be a bug but not seen anyone talking about 7990 issues not many have the card, can you do a visual of the card when its running any issues with fan stopping on that side of the card?
May 22, 2014, 11:20:52 AM
LOL (almost!)
Now it gets even weirder!!!
with your simple config, it had gpu0 OFF so I added "0,1" to the "device": line
and guess what?
Now gpu0 goes nuclear while gpu1 stays within limits!
Shortly after this gpu0 hit 98 so, back to myr-gr for a while.
(pls remember this is not 2 cards, it's 2 gpus on 1 card)
FEKKIN' COMPUTERS
Now it gets even weirder!!!
with your simple config, it had gpu0 OFF so I added "0,1" to the "device": line
and guess what?
Now gpu0 goes nuclear while gpu1 stays within limits!
Shortly after this gpu0 hit 98 so, back to myr-gr for a while.
(pls remember this is not 2 cards, it's 2 gpus on 1 card)
FEKKIN' COMPUTERS
May 22, 2014, 11:04:04 AM
Try to go as low with your memory freq as possible, you dont need high value here.
Maybe you could try to optimize your config starting from scratch (only basic settings) adding more options as you go. People are getting lower power usage and temperatures mining blakecoin than for example scrypt, so its definitely an issue on your end. Get correct settings for 1 card, and then add the other one.
Maybe you could try to optimize your config starting from scratch (only basic settings) adding more options as you go. People are getting lower power usage and temperatures mining blakecoin than for example scrypt, so its definitely an issue on your end. Get correct settings for 1 card, and then add the other one.
Code:
{
"pools": [
{
"url": "stratum+tcp://eu3.blakecoin.com:3334",
"user": "",
"pass": ""
}
],
"blake256": true,
"gpu-reorder": true,
"device": "1",
"intensity": "9,9",
"gpu-threads": "2",
"gpu-fan": "0-100",
"auto-fan" : true,
"temp-cutoff": "95",
"temp-overheat": "85",
"temp-target": "75",
"temp-hysteresis": "3"
}
"pools": [
{
"url": "stratum+tcp://eu3.blakecoin.com:3334",
"user": "",
"pass": ""
}
],
"blake256": true,
"gpu-reorder": true,
"device": "1",
"intensity": "9,9",
"gpu-threads": "2",
"gpu-fan": "0-100",
"auto-fan" : true,
"temp-cutoff": "95",
"temp-overheat": "85",
"temp-target": "75",
"temp-hysteresis": "3"
}
May 22, 2014, 10:49:50 AM
Even weirder - after tweaking as before, I used 'settings' to re-write the .config but when I re-ran it I still got the load error - but this time I was ready & screenied it 
Here's the config it wrote for me -
Here's the config it wrote for me -
Code:
{
"pools" : [
{
"url" : "stratum+tcp://eu3.blakecoin.com:3334",
"user" : "DiracConvertUK.beast",
"pass" : "x"
}
]
,
"intensity" : "20,3",
"vectors" : "1,1",
"worksize" : "256,256",
"kernel" : "blake256,blake256",
"gpu-engine" : "0-1100,0-1100",
"gpu-fan" : "30-100,30-100",
"gpu-memclock" : "800,800",
"gpu-memdiff" : "0,0",
"gpu-powertune" : "0,0",
"gpu-vddc" : "0.000,0.000",
"temp-cutoff" : "95,95",
"temp-overheat" : "90,90",
"temp-target" : "75,75",
"api-mcast-port" : "4028",
"api-port" : "4028",
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"log" : "5",
"no-pool-disable" : true,
"no-restart" : true,
"no-submit-stale" : true,
"queue" : "3",
"scan-time" : "30",
"blake256" : true,
"temp-hysteresis" : "4",
"shares" : "0",
"kernel-path" : "/"
}
"pools" : [
{
"url" : "stratum+tcp://eu3.blakecoin.com:3334",
"user" : "DiracConvertUK.beast",
"pass" : "x"
}
]
,
"intensity" : "20,3",
"vectors" : "1,1",
"worksize" : "256,256",
"kernel" : "blake256,blake256",
"gpu-engine" : "0-1100,0-1100",
"gpu-fan" : "30-100,30-100",
"gpu-memclock" : "800,800",
"gpu-memdiff" : "0,0",
"gpu-powertune" : "0,0",
"gpu-vddc" : "0.000,0.000",
"temp-cutoff" : "95,95",
"temp-overheat" : "90,90",
"temp-target" : "75,75",
"api-mcast-port" : "4028",
"api-port" : "4028",
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"log" : "5",
"no-pool-disable" : true,
"no-restart" : true,
"no-submit-stale" : true,
"queue" : "3",
"scan-time" : "30",
"blake256" : true,
"temp-hysteresis" : "4",
"shares" : "0",
"kernel-path" : "/"
}
May 22, 2014, 10:39:11 AM
Additional -
Another screenie, after tweaking config on the fly:-
To be clear, I have no problem with the behaviour of gpu0, I could run that a lot harder if only gpu1 would behave!
Cgminer won't even run using a .bat file it disappears before loading up
Another screenie, after tweaking config on the fly:-
To be clear, I have no problem with the behaviour of gpu0, I could run that a lot harder if only gpu1 would behave!
Cgminer won't even run using a .bat file it disappears before loading up
Code:
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
cgminer --blake256 -o stratum+tcp://eu3.blakecoin.com:3334 -u
DiracConvertUK.beast -p x -i 20,1 -g 1 -w 256 --thread-concurrency 24000
--gpu-engine 950 --gpu-memclock 1200
setx GPU_USE_SYNC_OBJECTS 1
cgminer --blake256 -o stratum+tcp://eu3.blakecoin.com:3334 -u
DiracConvertUK.beast -p x -i 20,1 -g 1 -w 256 --thread-concurrency 24000
--gpu-engine 950 --gpu-memclock 1200
May 22, 2014, 10:15:12 AM
7990 should give you more than 5GH/s, you may want to look at the posts in the beginning or middle part of this thread for a better config.
What core and memory frequencies are you using? And how many gpu-threads?
What core and memory frequencies are you using? And how many gpu-threads?
That's a LOT of reading, I just haven't got the time (with all my other 'duties'
) to do that.Core & Mem are pretty standard - 850/1200 on afterburner, fans set 100%. I reduced threads to 2 a couple of days ago.
BTW temps I posted for myr-gr were still on the cool-down from blake, right now it's 56/60 running 20Mh/s.
@mundivalur - cardS
An HD7990 has 2 gpus on one card, no riser. Even in my early days running scrypt I could hold max temp to 88 on gpu1.To me, the fact that gpu1 goes so hot so fast even with I=0 says there's a prob with the miner, blake is supposed to be more efficient than scrypt isn't it?
Here's the last config I tried, from last night - when I run it I get a (too fast to read properly) line saying that there is a prob in the config & it only partially loads.
Code:
{
"pools" : [
{
"url" : "stratum+tcp://eu3.blakecoin.com:3334",
"user" : "DiracConvertUK.beast",
"pass" : "x"
}
]
,
"intensity" : "20,1",
"vectors" : "1",
"worksize" : "256",
"kernel" : "blake256",
"gpu-engine" : "0-1100",
"gpu-fan" : "30-100",
"gpu-memclock" : "800",
"gpu-memdiff" : "0",
"gpu-powertune" : "0",
"gpu-vddc" : "0.000",
"temp-cutoff" : "95",
"temp-overheat" : "90",
"temp-target" : "75",
"api-port" : "4028",
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "2,1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"no-restart" : true,
"no-submit-stale" : true,
"queue" : "3",
"scan-time" : "30",
"blake256" : true,
"temp-hysteresis" : "4",
"shares" : "0",
"kernel-path" : "/"
}
"pools" : [
{
"url" : "stratum+tcp://eu3.blakecoin.com:3334",
"user" : "DiracConvertUK.beast",
"pass" : "x"
}
]
,
"intensity" : "20,1",
"vectors" : "1",
"worksize" : "256",
"kernel" : "blake256",
"gpu-engine" : "0-1100",
"gpu-fan" : "30-100",
"gpu-memclock" : "800",
"gpu-memdiff" : "0",
"gpu-powertune" : "0",
"gpu-vddc" : "0.000",
"temp-cutoff" : "95",
"temp-overheat" : "90",
"temp-target" : "75",
"api-port" : "4028",
"auto-gpu" : true,
"expiry" : "120",
"failover-only" : true,
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "2,1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"no-restart" : true,
"no-submit-stale" : true,
"queue" : "3",
"scan-time" : "30",
"blake256" : true,
"temp-hysteresis" : "4",
"shares" : "0",
"kernel-path" : "/"
}
May 22, 2014, 09:52:00 AM
Hey Guys, me again.
I REALLY want to luv this algo but it's driving me bats
No matter how I tweak my .config I get thermal runaway on gpu1 of my HD7990
This is one attempt, with I= 11,3, just before I turned off....
Then later, with gpu1 DISABLED & I=0 I got this before shutting down & going back to myr-gr
With the same rig this is how myr-gr behaves, so it's NOT the rig or card.
Is anyone else using an HD7990 who can post a known-good .config for me to try?
I know gpu1 always runs a bit hotter than gpu0 but on blake it's chuffing ridiculous right now
Do you have risers for the cards or are they close together at least gpu 1 looks extra hot even on myr so you might want to check that ! and you could also change the config and put higher numbers in temp cut off and over heat http://prntscr.com/3linc1 I REALLY want to luv this algo but it's driving me bats
No matter how I tweak my .config I get thermal runaway on gpu1 of my HD7990
This is one attempt, with I= 11,3, just before I turned off....
Then later, with gpu1 DISABLED & I=0 I got this before shutting down & going back to myr-gr
With the same rig this is how myr-gr behaves, so it's NOT the rig or card.
Is anyone else using an HD7990 who can post a known-good .config for me to try?
I know gpu1 always runs a bit hotter than gpu0 but on blake it's chuffing ridiculous right now
Jump to: