Topic: Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux) - page 140. (Read 3839201 times)

There is no speed difference between 12.1 and 12.2

Latest version v12.2:

NOTES:

After upgrading to 12.2 it has run for 3 hours without restarting.

BUT, gpu1 is set to to reduce intensity at 84 (-tt 76  -ttli 80,84) and I have found in the log where it has hit 90. The intensity is reduced (hash goes from 172 to 123 or lower) and the temperature slowly comes down.  I will try reducing the -ttli since there seems to be an over temp condition allowed.

20:05:00:931   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:00:934   43246780   GPU 6 temp = 88, old fan speed = 100%, new fan speed = 100%
20:05:00:935   43246780   Gpu 1 -ttli apply delta 4
20:05:03:937   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:03:940   43246780   GPU 6 temp = 85, old fan speed = 100%, new fan speed = 100%
20:05:03:940   43246780   Gpu 1 -ttli apply delta 1
20:05:06:942   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:06:945   43246780   GPU 6 temp = 89, old fan speed = 100%, new fan speed = 100%
20:05:06:945   43246780   Gpu 1 -ttli apply delta 5
20:05:09:947   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:09:950   43246780   GPU 6 temp = 85, old fan speed = 100%, new fan speed = 100%
20:05:09:951   43246780   Gpu 1 -ttli apply delta 1
20:05:12:953   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:12:956   43246780   GPU 6 temp = 90, old fan speed = 100%, new fan speed = 100%
20:05:12:956   43246780   Gpu 1 -ttli apply delta 6
20:05:15:958   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:15:961   43246780   GPU 6 temp = 86, old fan speed = 100%, new fan speed = 100%
20:05:15:961   43246780   Gpu 1 -ttli apply delta 2
20:05:18:963   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:18:967   43246780   GPU 6 temp = 87, old fan speed = 100%, new fan speed = 100%
20:05:18:967   43246780   Gpu 1 -ttli apply delta 3
20:05:21:969   43246780   GPU 0 temp = 75, old fan speed = 100%, new fan speed = 100%
20:05:21:973   43246780   GPU 6 temp = 87, old fan speed = 100%, new fan speed = 100%
20:05:21:973   43246780   Gpu 1 -ttli apply delta 3

How much speed increase has happened?Thanks for the update.

I may have some extras I would sell

That will do it, thanks!

i saw at least two people with the same problem of stability with 280x...but my idea is an option of setting intensity by the type of the vga, not the order,since that changes after every reboot...

Why -i3? My 280X with modded bios works fine on -i 5...

Claymore's Dual Miner Remote Monitor Android App can be used without wifi?

lol....already doing it smart guy...just giving an ideia or option to set intensity by the type of card, since mixed rig suffer after the reboot, because of the change of the order....but your ideia is very good to decrease the claymore profits...

Thanks for the update. Let's see if everything works fine now.

Claymore could you set -i 3 default for 280x?

Does anyone have a link to good reliable risers it seem 9 out 10 are crap

If you read next my posts, you will see that i found problem - it was in ttli option. It was bug in 12.1 version - when miner must to low intensity on achieving ttli temperature it causes GPU to hang and restarting.

In 12.2 all working normal as must.

P.S. I don't use modedd bios maked somebody. I maked it myself using my original ASUS Radeon R9 280X DirectCU II TOP (R9280X-DC2T-3GD5) bios.

Ramp up voltage on that card a bit up ... see what will happen, it should work normal with -i 5

Also, take off the extender cables, clean contacts on both card and small board that goes in motherboard. Re-test.

and ... last but not least - you mentioned "modded" 280x ? Which way it was modded ? You flashed some modded bios,
that you found on this forum (bad idea), or you manually modded lower straps into higher ones ?
also, what is the card manufacturer / exact model ?


If you mean clock/voltages ... gpu 1200-1233 , ram 1750, voltage, depending on asic quality from -80 to -96, memory straps from 1500MHz.


Of course, whan card is in slot, and working at higher PCI-E ratio/generation, cpu-to-gpu latency is lower, and thats why you see a bit higher speed.


b.r.
Alex

There is no speed difference between 12.2 and 12.1.

Correct me if I'm wrong... there is no speed increase in v12.1 or v12.2, right?

Or should I shift my rigs to the new version? Till now, I've been sticking with v12, and its all going smoothly.

https://i.gyazo.com/79fade7580b462b55180ea7cc6526be1.png
12.1 - 12.2
ubuntu 16.04
hashrate very low 290X 50kh/s
zec CPU load 100%

On my 280X all 12.x versions maximum 5 intensity...
On 12.1 with -i 5 - 2672Mb memory is used on 280X. On 270X with 4Gb and -i 6 3+Gb used. Therefore on 280X realy not enough memory to run over 5th intensity...

AFAIK this behavior is related to the available amount of GPU memory which limits possible intensity levels. Main/virtual mem should not be your problem here, if I am not totally wrong. Claymore stated that the memory requirements for the newer, faster kernels are higher.

On 12.0, .1, .2 running 5x280x with virtual memory of 56.6gb I'm getting
         GPU #3 is going to use too high intensity (6), not enough memory, intensity value reduced (5)
when I set just one card to -i 6 also happens with -old 1