X16R - RVN - Miner head to head test log

[opus.224]

Sorry for polluting your thread, JackIT, but I've made 2 tests with 2 identical rigs that I want share.
I wanted to know if mem underclock (-502) would underperform mem 0 performance.
Results looks interesting.
Rigs have same mobo, cpu, ram, ssd, psu, risers and 6 x ASUS STRIX 1070TI.
Tests were done with Enemy 1.09b and last exactly 24 hours.
All cards are set: 70% Power limit (126W), 70C Temp Limit, Core Clock: +100

1. test:
RIG1: ASUS_L: Memory Clock: -502, 291.49613601 RVN
RIG2: ASUS_R: Memory Clock: 0,  293.82484280 RVN

After 1. test RIG2 with higher mem clock gets advantage of 0.007988, or less than 1%.

2. test:
RIG1: ASUS_L: Memory Clock: 0, 287.17642980 RVN
RIG2: ASUS_R: Memory Clock: -502, 295.71330164 RVN

Looks like RIG2 is even more faster with mem clock -502.
Advantage of 0.029727, or 2.97%.
RIG2 looks faster despite of identical hardware.

Conclusion: underclocking mem could gain you slightly better reward.
 

[1715109332]

1715109332

[1715109332]

1715109332

[potificate]

Conclusion: underclocking mem could gain you slightly better reward.

With that power limit in place, this would make sense. Since mem adds to the power load, underclocking the memory would have the core throttled later. Have you tried this same test with higher limits (100% or higher)?

[opus.224]

Conclusion: underclocking mem could gain you slightly better reward.

With that power limit in place, this would make sense. Since mem adds to the power load, underclocking the memory would have the core throttled later. Have you tried this same test with higher limits (100% or higher)?

I'm limited with 1000W PSUs.
I could try a test with 80% PL maximum.
 

[potificate]

I'm limited with 1000W PSUs.
I could try a test with 80% PL maximum.

That would be great! Also, whatever temp limit you might feel comfortable with might help too.

[djn053]

any test for 1.10？

[JackIT]

any test for 1.10？

nothing yet.. haven't been mining RVN the past week.  as soon as come back to it, I'll run another test

[WickedPigeon]

any test for 1.10？

nothing yet.. haven't been mining RVN the past week.  as soon as come back to it, I'll run another test

What are you Mining? (If I may be so bold as to ask)

[vanscrew]

JackIT, do you have performed tests based on different components of the GPUs?

I want to know the result on how performs one and the same technology but with different specifications.
And if anyone can perform such test, that would be very helpful.

2 GPUs, with different specifications:

                         Dual GTX1060-O3G                     Dual GTX1060-O6G
GPU/Boost:        1594MHz/1809MHz                      1594MHz/1809MHz
RAM:                 3GB GDDR5/192bit 8008MHz        6GB GDDR5/192bit 8008MHz

* Obviously, the GPU/Boost is the same for both GPUs, would that mean same speed of mining?
* If all parameters are identical with exception of the RAM (3GB & 6GB), how would that reflect to the mining?

As of memory is not really a bottleneck for X16r, it shouldn't be noticeable.
There is no DAG things in X16r so memory capacity doesn't impact performance (like it does/did with ETH).
You will probably see different result if you cannot apply same OC on both cards (due to opposite ASIC quality).

Great! Thanks for the info, it is very helpful.
I have to read more about the algos.

[burster51]

Hi,

anyone has had a chance to test enemy 1.10 against suprminer 1.6?
I have been watching them for a few days, it seems that enemy is slightly ahead, just a feeling, but no solid testing my side.

Has anyone tried the performance of the other implementations e.g. xevan etc?

Thanks and nice thread!

[stevascha]

any test for 1.10？

nothing yet.. haven't been mining RVN the past week.  as soon as come back to it, I'll run another test

looking another testing here, nice posting man!