Noticeable hashrate drop across all rigs after new DAG last night?

MrYukonC

I don't know if it has always been this way and I just never noticed...

I woke up this morning and noticed that across all of my rigs there was a noticeable drop in hashrate of about 1+ Mh/s per rig.

All AMD cards -- 280X's & 370's.

Anyone else notice this?

ethfan

That is the effect of the increased DAG size. That's the reason Tahiti's have dropped from around 25 Mhps in the beginning to around 20 Mhps now.

MrYukonC

ethfan said:

That is the effect of the increased DAG size. That's the reason Tahiti's have dropped from around 25 Mhps in the beginning to around 20 Mhps now.

@ethfan Yeah, I figured it was just that, but never really noticed it like this before. I guess I need to pay more attention.

agent412

i woke up and my ethminer was closed, has this happened to anyone? This is the 3rd time this happened in the last 3 day. Is it because of the new DAG file? It's never happened before

hasher

My rigs are having issues too. This is really annoying.

MrYukonC

hasher said:

My rigs are having issues too. This is really annoying.

@hasher

I should also add -- my observations are not based on pool reports, but actual calculated hashrate output by ethminer.

I just find it odd that such a noticeable drop occurred with one DAG change. In fact, now that I think about it more, is it really possible that it dropped this much every DAG change for the past 6 months?

How many DAGs have there been since day 1? Every 30k blocks, right?

If so, that's ~43 DAG files.

~1 Mh/s drop across a 6 card rig last night is about ~0.1666 Mh/s per card.

0.1666 * 43 = 7.1666 Mh/s loss per card since day 1? That seems high.

280X's started out hashing at ~25+ Mh/s and are now down to 20-21 -- not 18-19 which is where that math would put them.

So yeah, something seems to be slightly up.

Bueller?

Genoil

Try running this https://github.com/Genoil/dagSimCL

Use same size for chunk and max to simulate current ethminer DAG implementation. You also need all the setx / export stuff ethminer requires these days. Batch files may not be fully up to date, I made this months ago.

MrYukonC

@Genoil Thanks. I'll give it a try at some point soon. Will probably run it on my single-card miner first since I can't afford the downtime on the larger 6-card rigs.

Genoil

It should give some insight into what's happening and what's coming up in the coming months. I'll try it on my new old 7950 too

ethfan

I see another drop in hash rate with yesterday's new DAG.

MrYukonC

ethfan said:

I see another drop in hash rate with yesterday's new DAG.

@ethfan Same here.

dlehenky

@MrYukonC How much are you dropping?

MrYukonC

dlehenky said:

@MrYukonC How much are you dropping?

@dlehenky About 0.5 - 1.0 Mh/s per rig -- and these are 6-card rigs that average(d) anywhere from from 90+ to 124+ (as of the previous DAG) -- now down to <90 - <123.

ethfan

Dropped approximately 0.7%

Shall we apply "CAGR" to this? At least this is much more predictable than mining difficulty or ETH pricing.

leotau12

For some reason my rig DL a DAG file 000000 something...???? Is this normal or what?

dlehenky

@leotau12 That's the genesis DAG that's created when you do a benchmark (ethminer -M).

leotau12

dlehenky said:

@leotau12 That's the genesis DAG that's created when you do a benchmark (ethminer -M).

OK?!?!?!?!?!

But I did not run benchmark command. Just noticed when I checked window to see if everything is running well.

dlehenky

@leotau12 Then I guess I don't understand what you are talking about. You *never* ran the benchmark? What window did you check? Where did you see it?

MrYukonC

dlehenky said:

@leotau12 Then I guess I don't understand what you are talking about. You *never* ran the benchmark? What window did you check? Where did you see it?

@dlehenky cc: @leotau12
He may be using Qtminer, which pretty much every freakin' time you stop it and restart it, it regenerates "a" DAG--not sure if it's the genesis DAG or the next one up, though.

leotau12

MrYukonC said:

dlehenky said:

@leotau12 Then I guess I don't understand what you are talking about. You *never* ran the benchmark? What window did you check? Where did you see it?

@dlehenky cc: @leotau12
He may be using Qtminer, which pretty much every freakin' time you stop it and restart it, it regenerates "a" DAG--not sure if it's the genesis DAG or the next one up, though.

Everytime I got home from work, I like watching my miners mine and check if something is going on (it's like watching TV) for some reason out of no where, it just created a DAG file *full-R23-0000000000000000* size: .99gb. Stopped my QTminer and deleted the file then restarted my QTminer.

Hope this sheds some light ^_^

dlehenky

@MrYukonC Yes, that's one of the reasons I don't use it. He's not giving me enough information to know what he's talking about.

dlehenky

@leotau12 As @MrYukonC said, qtminer always creates the genesis DAG, which is totally useless, but that's what it does. There's no harm in it, other than wasted startup time. Stop deleting it. You're just causing it to repeatedly rebuild the DAGs. Leave it alone.

leotau12

Sorry ^_^

First time I have ever seen it after a week of mining.

dlehenky

@leotau12 No worries. I'm just letting you know you're causing yourself problems by deleting it, if your running qtminer. Normally that DAG only gets generated when you run a benchmark, if it isn't already built.

Genoil

Well this is approximately where it's going with Tahiti cards:



It's the output of my DAGGER simulator. For soem reason it won't run on my 970 any longer, but that one kind keeps a steady 20MH/s until 2GB and then it takes a steep nose dive. Kind of like the 780 (old pic taken from thread):

ethfan

Just a little while more for the next DAG to hit the hash rate...

Biodom

@Genoil
Interesting graphs. When you post 7950, do you mean 2gb or 3gb cards?
If the latter, then why would 3gb cards be so negatively affected starting at 1.5 gb DAG.
If the former, then I suggest to make a clarification.

Genoil

@Biodom it's a 3GB card and it started at 1024MB, the DAG size at epoch 0. A 2GB card will show the same curve, but it would just abruptly end at 2048MB. It doesn't have to do with the size of the RAM but with the TLB. (translation lookaside buffer). That's a unit in the IOMMU (memory management unit) that translates virtual memory addresses (as used by the OS) into physical memory addresses (as used by the hardware). The TLB caches lookups to specific regions (pages) of the physical memory in its tables, so that when another address from that region is requested, it can quickly give the IOMMU an answer.

For normal GPU operations that TLB is large enough to cache most requested addresses, but the problem with ethash is that it does pseudo-random lookups accross a very large memory range. As the size of the DAG increases, the TLB runs full quicker, causing more memory lookups to have to "replayed". This is called TLB trashing.

As you can see both AMD and Nvidia cards suffer from it, but in a different way. Between Compute and GCN versions there's also a difference. I haven't seen too many results of people with GCN 1.1 and 1.2, would be interesting to see.

Biodom

@Genoil
I am still a bit unclear about which graph to follow for 7950-I assume upper and not lower. However, in this case simulation is not exactly correct because so far i did not observe any difference in MSI HD7950 card (at steady 19.3Mh/s) during the last 1.5 mo while on the graph it should be decreasing steadily on upper picture. If it is the lower pict, then it is different, of course.
At https://forum.ethereum.org/discussion/3947/dagger-simulator
one guy (chinhdn) had done 7950 stimulation and he has nothing going wrong at 256mb chunks until DAG size 2560 mb.
Does the single chunk alloc command in Linux:
export GPU_MAX_ALLOC_PERCENT=100
export GPU_SINGLE_ALLOC_PERCENT=100
that people use now basically negates the problem or exacerbates it?
Thanks for the prior explanation; could you comment further?
I would appreciate it very much.

Genoil

It's the upper graph, the lower is for GTX780. I see my 7950 perform a little bit worse with every new DAG.

The chunking strategy I had been working on was bugged. I fixed it the other day and it has a negative impact on hashrate. So it's quite useless even more so now people have found the right environment vars to set.

curt162

Is there any more news about this? When the last Dag update occurred I lost about 1Mh/s PER CARD.
I can't imagine it's just the dag, but I cant regain the lost hashing power.