Topic: Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux) - page 251. (Read 3839204 times)
I changed and 0..3, increased the voltage across avterburner not help hash is less than 1 - 2%?
Nice improvements 
For all my cards I get 4% improvement. I got it for undervolted and overlocked 390 with -a 1, and 280x with -a 1 or -a 2 depending on the model. For example XFX are better with -a 2 (215 sols vs 210 with -a 1), while Saphire Tri x is way better with -a 1, where I get 230 sols on stock cards vs 218 with -a 2.
v10.0 only decreases hash r9 390 yf 1-2%.who took the decision?
What does Cannot get Overdrive capabilities for adapter in logs mean?
Means your gpu is not stable, increase the voltage or lower your clocks, or check for other errors
v10.0:
- improved speed for most cards, up to 8% (depends on card model). About 340H/s on stock 390X, 255H/s on stock RX480, 320H/x on stock Nano.
- added "-a" option to select best algorithm implementation for your GPU.
PS. This is an intermediate version because I had no time to implement all optimizations I planned for this release, also I had no time for tests. In 1-3 days I will release another version with further speedup.
- improved speed for most cards, up to 8% (depends on card model). About 340H/s on stock 390X, 255H/s on stock RX480, 320H/x on stock Nano.
- added "-a" option to select best algorithm implementation for your GPU.
PS. This is an intermediate version because I had no time to implement all optimizations I planned for this release, also I had no time for tests. In 1-3 days I will release another version with further speedup.
240 sols on my 7950. simply wow. Have to try the 6970
v10.0:
- improved speed for most cards, up to 8% (depends on card model). About 340H/s on stock 390X, 255H/s on stock RX480, 320H/x on stock Nano.
- added "-a" option to select best algorithm implementation for your GPU.
PS. This is an intermediate version because I had no time to implement all optimizations I planned for this release, also I had no time for tests. In 1-3 days I will release another version with further speedup.
- improved speed for most cards, up to 8% (depends on card model). About 340H/s on stock 390X, 255H/s on stock RX480, 320H/x on stock Nano.
- added "-a" option to select best algorithm implementation for your GPU.
PS. This is an intermediate version because I had no time to implement all optimizations I planned for this release, also I had no time for tests. In 1-3 days I will release another version with further speedup.
240 sols on my 7950. simply wow. Have to try the 6970
What does Cannot get Overdrive capabilities for adapter in logs mean?
RX 480 a2 260h/s
R9 Nano a1 350h/s
R9 Fury a2 375h/s
Not sure why the R9 Nano performs better on the algorithm 1.
The R9 Fury has 56 compute units, while the R9 Nano has 64.
In previous versions the R9 Nano has always outperformed the R9 Fury.
I think some more optimization can probably be done on the R9 Nano for v10.
Nice speed increase overall, thanks Claymore!
How to increase hashrate in r9 290 saphire?
claymore v9.3 r9 290- 1050/1500 give only 295-298 h/s with overclock.
How can i get more?
P.S. v10.0 only decreases hash on r9 290 - 290 h/s.
claymore v9.3 r9 290- 1050/1500 give only 295-298 h/s with overclock.
How can i get more?
P.S. v10.0 only decreases hash on r9 290 - 290 h/s.
V10.0 is much more stable but uses more power.
Can u be more specific how much exactly with examples,thx in advance
V10.0 is much more stable but uses more power.
r9 280x has about 5-6% improvement, r9 390 about 2% slower.
9.3 390 - 313 sol
10.0 390 - 305 sol
is bad
10.0 390 - 305 sol
is bad
Yes but not that much drop
rx 480 8 gb
Hi,
-a 2 or -a 3 select algorithm 2
-a1 select algorithm 1
-a 2 or -a 3 select algorithm 2
-a1 select algorithm 1
-a 2 is default/max Algo
you can set -a 1 or -a 2
if not then remain it on auto
it will detect best algo automatically best fit for your GPU
Gigabyte G1 rx 470 4 GB, memory 1810, tuned timings, i8, -a 2.
1220 - 242 h/s
1270 - 251 h/s
1350 - 260 h/s
It's average values from EthMan.
Working mode is 0.98 V and 1220-1270 MHz depends on card.
With 9.2 i had 229 h/s on 1270 MHz, so about 10% increase.
Power consumption increased about 2-3%.
1220 - 242 h/s
1270 - 251 h/s
1350 - 260 h/s
It's average values from EthMan.
Working mode is 0.98 V and 1220-1270 MHz depends on card.
With 9.2 i had 229 h/s on 1270 MHz, so about 10% increase.
Power consumption increased about 2-3%.
well, i8 -96mv crashes on site,did 364 before. Will try stock.
yeah, all fiji gpus(20+) have a minus 20 hr, plus freeze on 3 rigs at i8.
Which driver version are you using?
What is your -a set to?
I have Furys that went from 345 H/s to 375 H/s (v9.3 to v10.0).
I have Win10 running 16.11.5 drivers with -a 2
Edit: I run -i 6 using stock bios and all stock clocks. Fijis are Sapphire Fury Nitro.
280 1000/1500 -i 5 from 195 to 210
290 1050/1250 -i 7 from 320 to 320 or slightly slower
-a 0 works fine
-a 3 not at all
290 1050/1250 -i 7 from 320 to 320 or slightly slower
-a 0 works fine
-a 3 not at all
Your GPU3 is a RX480 with 36 compute units. The RX470 has only 32 compute units.
As i said before, up to 270h/s (and beyond) from single rx470 4gb
The only thing I see is that you pushed it to hard and GPU0 is failing - so I would not be bragging about it
Jump to: