Topic: [ANN] TeamRedMiner v0.10.10 - Ironfish/Kaspa/ZIL/Kawpow/Etchash and More - page 114. (Read 211959 times)

Lyra2zz implementation: https://github.com/fancyIX/sgminer-phi2-branch/commit/3054ddaeadad5a3797f611872f4b298052a2273f
 

GRIN, BEAM or XZC would be awesome to add. We all need some more profit these days

hi dev Can you pretty please add support for r cards too (r280, 290 etc)? your miner is epic! and soooo stable. getting almost 1000h on xmr on my rx 480 and 570

Any plan to support GRIN-29/31?

Hi TRM,

any plan to support MTP algorithm?

Thanks.

TRM is the best v8 miner for my rig, highest hashrates and lowest power draw.
It seems to be more aggressive on memory compared to SRB and JCE, on some cards I had to underclock memory and increase memory voltage to avoid errors.

Are you considering Stellite?  It just forked this week so it would be a great time to add it.

Thanks for the input guys !!
Final verdict : TRM rocks, NH sucks !!

My Hashrate and stats are about the same as the screenshot posted about 2050 H/s with about 4% rejections !!

Thx guys, bye bye NH   

The power draw by this miner varies a lot over time as we move through the different phases of the CN algo. The Vegas have their Adaptive Voltage and Frequency Scaling (AVFS), which is what you're seeing in action in a very direct way in the video. Afaik, Polaris cards does not have any dynamic realtime scaling, they'll just throttle you in case of a thermal of power limit breach. I'm not an expert in this field though, others here are more knowledgeable, so if you want to know more I'm sure someone else can pitch in.

Are those variations of voltage and core clocks normal in my vega?
My rx 570 is pretty stable beside

https://youtu.be/MC4owxoYx9g

Thank you!

GOOD LUCK MINING--

I have had a good run with the latest boot of my 4X RX VEGA 64 rig.  The calculated, average, and poolside hash rates are all very close.  I thought I'd post a screen cap:

TeamRedMiner --

This is the best performance on the Vega rig with TRM v0.3.8 since downloading it.  There is only a 2.74% gap between calculated and poolside hash rates.  However, earlier runs have had much larger differences between calculated console and poolside hash rates, with the calculated poolside hash rate at or near 7.5kH/s, and higher reject rates.

The "--cn_config 15+15" has been the same for weeks.  Regardless, even at the lowest calculated poolside hash rate, the Vegas perform about 1/3 better than with XMR-Stak on the same Linux rig.

Looking forward to the next version!       --scryptr     

Hey!

Yeah, it's high time to ramp up TRM dev work again now, we've had some time off working on other projects for 1+ month now. I think there are lots of interesting work to be done. We might very well venture into the equihash and cuckoo cycle type of algos, but we have some things to wrap up before that: vertcoin fork, finally add some more CN variants, add some more features to the miner itself.

But yes, we will def make good on our promise to bring AMD up to par with Nvidia whenever the raw hashrate difference is too large. Efficiency is a different thing though. Furthermore, the algo(s) need to be large enough to make it worthwhile for us to spend the time.

Stay tuned though, some interesting stuff will appear within a few weeks .

I tried multiple times with different miners and connecting to nicehash for CN sucks constant disconnects and rejects I sent multiple request to support about this for almost 2 years now not to the amount of other miners with the same trouble but for some reason they won't fix it

@trm devs

Hey, not sure if this is where you would like questions like this to be posed as it not CN related
Anyway, It sounded like your initial plans was to bring to agility to AMD cards where the Nvidia cards normally dominated

Just would like to know if you plan on bringing new algos to the party. Especially for example at the moment BEAM coin seem to be doing really good and the performance of the initial miners for AMD seems to be quite good. The Vegas were running around the same speed as the 1080Tis at the start. However, the guys that normally do dev for nvidia cards have picked up algo and have brought some big improvements with promises of more

Knowing your team has insane ability of producing serious hashrates with the AMD cards
Just want to know if you are at all considering bringing out your miners for BEAM (Equihash 150_5) and GRIN

EDIT: not too sure what algo GRIN is using

I've written about this a number of times . Nicehash is a bitch to use for GPUs mining CN. New jobs can be sent out as often as every 5 secs during extended periods, and your effective hashrate is reduced considerably. The underlying reason is that computing a batch of hashes on the gpu takes 1-1.5 secs. If a new job arrives halfway through this computation, and the pool is very picky about accepting stale shares, you'll see the reject ratio you're having. Moreover, we currently try to submit all shares found in TRM, even though we know we've just received a new job.

In the general case, a pool should (imho) accept all shares that still would be possible to submit as a network block assuming they had the necessary share difficulty. Supportxmr and MPH does this, Nanopool is more picky. For Nicehash, it seems they don't accept any stale shares at all.

Castxmr and maybe other miners(?) have an abort mechanism, i.e. they just discard any ongoing work on the gpu when a new job arrives. We should add the same feature in some upcoming release. It's no silver bullet though, you still lose all the work you had done before the abort. It's on the TODO list for TRM, currently all you can do is benchmark the miners available, check your power draw, and the one that produces the best poolside hashrate for you over a week should be your weapon of choice. Nicehash is a bitch with variance though given how they enforce high difficulty, so it really is a pain to test things, that's why I recommend a full week above .

Change the pool to one closer to you. A 205 ms is too much. You are using the eu pool, try the usa pool see if it's the same....

Hey guys, Using TRM on nicehash causes many rejected shares about 4%.
Earlier used TRM on miningpoolhub and supportxmr with 0 rejects.

Driver : 18.6.1
Cards : Vega 64
Config : 15+15

This is the rejection :
Is there a specific parameter to enable the nicehash support?
On the same note is the "Keepalive" parameter supported?

Thx

Hey, sorry about the delayed reply. I'm not a straps expert myself, I've been lucky enough with one click mods for my own Polaris rigs. Your best hope is to google around or hope some other user with confirmed good CN straps for those Hynix modules chips in.

Yes I change my cards BIOS and strap is this for Hinyx memmory

HYNIX H5GQ8H24MJR
999000000000000022339D00CEDD6B4470551414B80D960B0040C400750314204A8900A00200000 018123745B92B3A16

Hi!

Since you’re seeing the same hashrate with all configs, you’re bound by mem bandwidth. You need to work with your mem straps. Have you done any bios mods for the straps? If so, what straps are you running?