Bro, can you tell me you settings about RX560 - v1.6.9 = 613H/s, my cards only 550H/s
Oтвeт нa твoй вoпpoc бyдeт нижe...
See answer for your question below...
Okay, so some of you said that the hashrate on Rx560 4g is lower on 1.7.0 vs 1.6.9 on heavy algos (said Haven) -
HEREI took some time and made some tests :
[overquoting skipped]
So as you can see i can't reproduce the huge -100hs drop you said. As a matter of fact, i can only produce + hash, not -

Something really strange going on. Made a few benchmarks...
The card and driver:

Memory is micron MT51J256M3, BIOS is flashed with this strap in all above the 1750:
777000000000000022AA1C00EF615C41B0551016BA0D96060060060004061420EA8940AA030004C
01914292E692E3B16
Every miner version has compiled kernels by it's own. Every test duration is 6+ min. ODT reset and re-apply settings before each miner start.
And all of these tests is on
haven algo. I think, for heavy will be the same config.
Parameters, used in config.txt:
i - intensity
w - worksize
t - threads
f - fragments
Core/mem clocks are 1200/2025
Results: i29w16t2f1
v1.6.9 - 512
v1.7.0 - 499
i29w16t2f32
v1.6.9 - 610
v1.7.0 - 505
i29w16t2f16
v1.6.9 - 526
v1.7.0 - 502
i29w8t2f1
v1.6.9 - 539
v1.7.0 - 502
i29w8t2f16
v1.6.9 - 612
v1.7.0 - 505
i29w8t2f32
v1.6.9 - 608
v1.7.0 - 506
i29w8t2f64
v1.6.9 - 606
v1.7.0 - 513
Question 1: how to calculate starting intensity to tune the card? Somewhere I've found a formula, that best intensity = (shaders-64)/2, but all of my cards are producing compute errors on such values, so for each card I need to go down for 128-256 points, to get rid of errors...
Question2: how the intensity value is correlated with vmem size? For example, if RX560 2Gb card have max hashrate at intensity 16, is it possible for RX560 4Gb card to have bigger hashrate at bigger intensity? And how much is "tuning window" for 4Gb model, comparing to 2Gb? (assuming, that # of shaders and CU are the same on both cards)
Question3: is there a way to set intensity per-thread for each card? So, I mean, for example, 2 threads per 1 card, but 1st thread with i54, and the 2nd thread with i56? (maybe feature request for future ver?..)
I will introduce a few modes for heavy algos in next version, so everyone can find the one that works for him best.
1. Optimal intensity can sometimes differ even for the same brand cards, so i don't really know how to calculate the best value. Unfortunately its a try-fail game until you find the best one. It's not always the highest intensity that gives the best results.
2. Intensity = number of global threads (in other miners that value is SRB intensity x 16). Now when you do scratchpad size * global threads you get the approx size of memory that will be used.
example for heavy : 56 * 16 * 4 = 3584mb.
3. There is, check the first page
(--enableduplicategpuid (enables the usage of the same gpu id in gpu_conf multiple times))