GPU load going bonkers

[kweejee]

I'm currently hashing with several GPUs. My issue is on my R9 290 Powercolor. The problem is that the GPU load goes all over the place. It ranges from 10% to 100% while hashing. It's literally all over the place. Sometimes its 15% then it shoots to 100% then back down to 55% then to 70% so on and so forth. And I cannot figure out why. The hashrate on said GPU also shoots everywhere. Sometimes its 860k then other times its 760k and everywhere in between.

Anyone have some tips or suggestions?

[kweejee]

Also the card in questions runs under these settings:

Worksize 256 TC 20481 lockup 0 GPU engine 1020 mem clock 1500 gpu threads 2 powetune 20 temp target 75

[djm34]

on what are you running ?
If scrypt/scrypt-n I would suspect that it is coming from the low temp target (temp target 75 and powertune 20 is not a good idea, either you increase the temp target, or you decrease the powertune)

[kweejee]

on what are you running ?
If scrypt/scrypt-n I would suspect that it is coming from the low temp target (temp target 75 and powertune 20 is not a good idea, either you increase the temp target, or you decrease the powertune)

I'll try a temp target of 80 and lower powertune to 15. I'm gunna try force constant voltage too. I saw my VDDC in jumping like a rabbit

EDIT: Still going bonkers. Removed powertune entirely too and tried it.

[hashrr.com]

I am running the r9 290 Powercolor, stable. Here are my settings.

"intensity" : "18",
"vectors" : "1",
"thread-concurrency" : "35960",
"worksize" : "256",
"lookup-gap" : "2",
"auto-fan" : true,
"gpu-engine" : "941",
"gpu-memclock" : "1250",
"gpu-powertune" : "20",
"sharethreads" : "48",
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"temp-cutoff" : "97",
"temp-overheat" : "88",
"temp-target" : "74",
"temp-hysteresis" : "3",
"queue" : "1",
"scan-time" : "60",
"scrypt" : true

[kweejee]

I am running the r9 290 Powercolor, stable. Here are my settings.

"intensity" : "18",
"vectors" : "1",
"thread-concurrency" : "35960",
"worksize" : "256",
"lookup-gap" : "2",
"auto-fan" : true,
"gpu-engine" : "941",
"gpu-memclock" : "1250",
"gpu-powertune" : "20",
"sharethreads" : "48",
"expiry" : "120",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "1",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"temp-cutoff" : "97",
"temp-overheat" : "88",
"temp-target" : "74",
"temp-hysteresis" : "3",
"queue" : "1",
"scan-time" : "60",
"scrypt" : true

I can't get my TC higher than 27400 without causing my GPU to shut down in sgminer. Even though I have 12 gigs of ram

[kweejee]

Bump

[hashrr.com]

What are your current settings? GPU-Threads, Memclock, Gpu-Engine?

[hashrr.com]

This is more than likely caused by not setting a constant fan speed. Newer bioses of certain cards have this issue.

I use temp target 74 on my r9-290's (I have many rigs). The GPU load remains constant at 100%. The fan's on the r9 just pickup speed when the temp goes above that, hashing remains the same. It is possible there isn't enough external cooling on the rig. I use blower fan's to keep the r9's cool and running at a constant temp/hash rate. In my experience though, to get a constant hash-rate on the r9's you need to get your settings just right.