Author

Topic: GPU Starting Slowly (Read 583 times)

full member
Activity: 163
Merit: 100
December 18, 2013, 09:08:36 AM
#8

I 12
TC 8192
W 256
G 2

GPU 0 runs at 69-70C consistently and GPU 1 runs at 60-62C. I have been running GPU 0 about 2 weeks longer as it was part of my initial PC build.


I have 2 of these cards that run stable at around 405kh/s with the following settings:

-I 20,19 -g 1 -w 256,256 --thread-concurrency 15296,15936 --gpu-engine 930,960 --gpu-memclock 1393,1438

Definitely run -g 1 and try increasing your thread concurrency significantly. This should allow you to increase your intensity. Also, make sure you set in Windows/Linux GPU MAX ALLOC PERCENT=100
hero member
Activity: 831
Merit: 500
BitSong is a decentralized music streaming platfor
December 18, 2013, 08:57:16 AM
#7
Thanks for the suggestions. I have been through the hardware comparison pages and some other sites with people giving their settings for the 7870 as well.

I have not tried G1 yet. I will give that a try. Both cards are still hashing away.
hero member
Activity: 532
Merit: 500
December 17, 2013, 11:23:36 PM
#6
Good guide for settings on various cards here:

https://litecoin.info/Mining_hardware_comparison
sr. member
Activity: 490
Merit: 254
December 17, 2013, 10:54:25 PM
#5
Have you tried higher thread concurrencies with G 1 and higher intensity, say 18?  I have a few cards myself that don't like G 2, but run well at G1 with high TC and intensity. Might want to do a search on your card to see if you can find some settings to try.
hero member
Activity: 831
Merit: 500
BitSong is a decentralized music streaming platfor
December 17, 2013, 10:50:44 PM
#4
I 13 took it down. I tried to go through all the settings with intensity, thread concurrency, work size and the engine and memory clocks through Catalyst.

I have a Sabertooth Z77 MB and GPU 0 is in the x16 slot closest to the CPU and GPU 1 is in the next PCIe 3.0 slot down (x8). Right now, they are at 70C and 57C with fan speeds of 2780 and 2034. Both are hashing normally, but GPU 0 has been taking a few seconds to get up to speed on start-up of CGMiner.
sr. member
Activity: 266
Merit: 250
December 17, 2013, 07:52:04 PM
#3
where are the cards positioned on the motherboard in relation to eachother?  the hotter one may be underperforming due to inadequate cooling
sr. member
Activity: 490
Merit: 254
December 17, 2013, 07:49:43 PM
#2
No luck at I 13?

I don't think you can read much into anything in the first 30 seconds, much of the time the information at this point is basically going to be guessed at anyway by the miner until it has enough of a baseline to bring the numbers inline with actually is happening. If they seem to be about even after 5 minutes I don't think you have much to worry about.
hero member
Activity: 831
Merit: 500
BitSong is a decentralized music streaming platfor
December 17, 2013, 07:08:23 PM
#1
Hey all-

I have (2) Sapphire 7870 GHz Editions and have been mining for a couple of weeks. Both GPUs are running at stock settings in an ATX Mid Tower set-up, cover on. I am scrypt mining with settings of

I 12
TC 8192
W 256
G 2

GPU 0 runs at 69-70C consistently and GPU 1 runs at 60-62C. I have been running GPU 0 about 2 weeks longer as it was part of my initial PC build.

In the last couple of days, GPU 0 has taken a little time to get up to speed. About 30sec to get up to speed. GPU 1 starts right up. I have an NZXT Hale 82 700w PSU.

Is this a sign of a dying GPU? Any insightis appriciated.

Thank you.
Jump to: