any miners like CGMiner able to switch to a pool when rejects on one go significantly above average?
cgminer switches if it encounters 3 minutes of continuous rejects |
|
|
|
[02:36] <@conman> stratum is basically orthogonal to all the hard work I put into network scheduling D: It doesn't mean stratum is a bad thing, it's just a shame that I put a lot of work into making cgminer scale to massive workloads and I have to implement something else from scratch again. |
|
|
|
...and it just stays like that forever, without submitting any work.
catalyst 12.8
AMD stream sdk 2.1
SDK 2.1 is incompatible with any catalyst >12.5 i uninstalled (in this order): amd stream sdk, catalyst 12.8, and installed 12.4 and amd stream sdk 2.1, but i'm still getting the same problem. After you change sdk you have to delete any .bin files in your cgminer directory for it to recreate them with the new sdk. Also I was guessing which catalyst version in particular it was, I know it's around then, but 11.12 is meant to be a better driver for it anyway. |
|
|
|
Is there even a normal zip of the new cgminer on here? I feel like throwing my computer through the wall every time I come here, for the new update.
Use 7-zip - it's free and you won't be breaking the winzip license like almost everyone does with winzip ... http://www.7-zip.org/Why would you be breaking the license? It's a free download. Installing a new compression software on a dozen computers for no purpose, is absurd. So wait, you're happy to go to great lengths to set up your hardware to mine, install special software for it to do so, drivers sdk etc and configure them, and then you're happy to download my free software from my website that I pay for for you to have free unrestricted access to, and you're complaining because you have to install 7z? It sounds so bad when you say it  . I prefer .arc by the way. Sending another BTC just because I haven't in a while. Thanks, Sam Thanks! Would you like a .zip uploaded for your donation?  |
|
|
|
Is there even a normal zip of the new cgminer on here? I feel like throwing my computer through the wall every time I come here, for the new update.
Use 7-zip - it's free and you won't be breaking the winzip license like almost everyone does with winzip ... http://www.7-zip.org/windows has had .zip support built in for ages. no need to use .7z. M Who the heck uses Winzip anymore?!?!? Supporters of bloatware I guess. The .zip implementation in Windoze doesn't do .7z format. 7Zip is a great archiver/compression utility and is cross platform. I would highly recommend it. Sam Don't get me wrong. I like 7zip's compression as much as any other geek. (I also think the UI has lots of room for improvement, which is irrelevant here.) But above all, not everyone is a techie. We're not talking much size difference put making a windows compatible .zip vs a .7z file. You can even use 7zip to make it if you want. Using windows .zip makes it more accessible to others. M I'm well aware and that's why I used to compress it with zip, but with bitstreams on fpgas making the archive so goddman large, it is an inconvenience and it does use up bandwidth when I'm not on a fast wire to the server itself. The 7z archive is about 1/3 the size of the zip. |
|
|
|
Is there even a normal zip of the new cgminer on here? I feel like throwing my computer through the wall every time I come here, for the new update.
Use 7-zip - it's free and you won't be breaking the winzip license like almost everyone does with winzip ... http://www.7-zip.org/Why would you be breaking the license? It's a free download. Installing a new compression software on a dozen computers for no purpose, is absurd. So wait, you're happy to go to great lengths to set up your hardware to mine, install special software for it to do so, drivers sdk etc and configure them, and then you're happy to download my free software from my website that I pay for for you to have free unrestricted access to, and you're complaining because you have to install 7z? |
|
|
|
|
I run 4x7970 overclocked at >2.7GH total on debian linux and everything else in the case with the one 1250W PSU and have done so for 6 months solid. Not all PSUs are created equal.
|
|
|
|
...and it just stays like that forever, without submitting any work.
catalyst 12.8
AMD stream sdk 2.1
SDK 2.1 is incompatible with any catalyst >12.5 |
|
|
|
Is the --net-delay option still useful with the latest version of cgminer?
Thanks,
gigavps
Still active in the README It still works, but the current cgminer is much nicer with connections all round anyway so I doubt anyone will require this option unless they have a router that only allows something like 8 concurrent connections. |
|
|
|
|
Looks good to me. Clearly any claim that jesusminer is better is pure fud.
|
|
|
|
sunbreak If i am informed correctly, sunbreak does not have a stake in this one way or the other, he would be a good impartial judge. He is also very knowledgeable in the area. edit: or me Sure, I'll go ... and it's true ... I do not own any fpga's or have anything invested in asic's nor did I invest any money with pirate for that matter. As for credentials; I personally run a 57 ghash/sec gpu only mining cluster on 110x6950's and 24x5970's on 24 machines with no AC and using only 20kW. I'm the author of strippedminer the 187mb netboot mining distro. I originally wrote all the --auto-fan --auto-gpu code that is in cgminer as an external script and independently sponsored the ADL implementation in cgminer by conman). I also was the one suffered lengthy testing with conman to reveal proper methods to restart stalled gpu's in cgminer. I identified and diagnosed the lack of dns caching and http connection reuse in cgminer which conman reimplemented as a curl ring buffer saving all of our routers Well that's enough, I've done lots of stuff in the community ... and I would love to get a chance to put bfl to the test. I'd be happy for sunbreak and abracadabra to go. After that... anyone but Jesus. |
|
|
|
No.
Damn... Hah! What makes you think I'd want to spend my own money on international airfares to benefit BFL? |
|
|
|
I don't think the estimation from the server will be perfect to give you exactly the right number of getworks. It will only ever be approximate within a range of timeframes. If you run cgminer 2.7.6 it will show you what diff shares you are submitting now.
I upgraded cgminer from 2.7.5 to 2.7.6 today and ran it for a few hours on both a regular server, and the dynamic test server to compare them. This is with a regular server (us3.eclipsemc.com:8337): cgminer version 2.7.6 - Started: [2012-09-25 13:50:55]
--------------------------------------------------------------------------------
(5s):1878.9 (avg):1596.6 Mh/s | Q:154 A:1979 R:6 HW:0 E:1285% U:15.9/m
TQ: 0 ST: 3 SS: 0 DW: 17 NB: 14 LW: 5007 GF: 1 RF: 0 WU: 22.1
Connected to http://us3.eclipsemc.com:8337 with LP as user ---
Block: 00000343360189b2e6653ae3f7eb361e... Started: [15:28:23]
--------------------------------------------------------------------------------
[P]ool management [S]ettings [D]isplay options [Q]uit
BFL 0: 59.8C | 803.7/798.9Mh/s | A:1000 R:4 HW:0 U: 8.06/m
BFL 1: 62.7C | 803.8/797.7Mh/s | A: 981 R:2 HW:0 U: 7.90/m
--------------------------------------------------------------------------------
[2012-09-25 15:52:31] Accepted 0e608cb9 Diff 1 BFL 0 pool 0
This is with the dynamic test server (us3.eclipsemc.com:8437): cgminer version 2.7.6 - Started: [2012-09-25 15:56:27]
--------------------------------------------------------------------------------
(5s):1468.6 (avg):1582.5 Mh/s | Q:238 A:1003 R:0 HW:0 E:421% U:8.3/m
TQ: 0 ST: 3 SS: 0 DW: 50 NB: 15 LW: 4656 GF: 1 RF: 0 WU: 22.1
Connected to http://us3.eclipsemc.com:8437 with LP as user ---
Block: 00000005f1bf5387d0157105a798cd0e... Started: [17:56:50]
--------------------------------------------------------------------------------
[P]ool management [S]ettings [D]isplay options [Q]uit
BFL 0: 59.8C | 677.6/791.6Mh/s | A:508 R:0 HW:0 U: 4.20/m
BFL 1: 62.6C | 675.1/791.6Mh/s | A:496 R:0 HW:0 U: 4.10/m
--------------------------------------------------------------------------------
[2012-09-25 17:53:19] Accepted 49cb55c7 Diff 3 BFL 0 pool 0
[2012-09-25 17:56:32] Accepted 41e5c544 Diff 4 BFL 0 pool 0
[2012-09-25 17:56:42] Accepted 3e5718c1 Diff 4 BFL 0 pool 0
[2012-09-25 17:56:50] LONGPOLL from pool 0 detected new block
[2012-09-25 17:57:05] LONGPOLL from pool 0 requested work restart
[2012-09-25 17:57:16] Accepted 062c4e21 Diff 4 BFL 0 pool 0
Looks perfect. Note how the WU is identical on both counts, and that the dynamic server is changing the diff on more than one occasion. It will start at diff 1 and slowly converge on the best balanced diff over time. Inaba, notice what I have reported before how there are 2 longpolls 15 seconds apart on dynamic diff though? cgminer can tell when a longpoll is associated with a new block or just a request for a work restart. This 2nd longpoll still seems wrong. |
|
|
|
|
因幡の白兎っていうの神話的作り話からの「いなば」でしょう
|
|
|
|
My first reaction was ckolivas. Maybe kano if he can't. My second opinion is Graet, but I wanna say he's in AU.
They're all in Australia. Yes, we're all in Australia, in various different cities. Well, do you want to go to Kansas City? (in the US preferably, unless the non-US person wants to pony up a ticket) No. |
|
|
|
My first reaction was ckolivas. Maybe kano if he can't. My second opinion is Graet, but I wanna say he's in AU.
They're all in Australia. Yes, we're all in Australia, in various different cities. |
|
|
|
I probably need to repackage the tarball. Gimme a few mins. I've reuploaded the source tarball for those having the ADL_SDK problem. Please try again. Worked for me. Thanks. It was a simple packaging error on my part. |
|
|
|
I probably need to repackage the tarball. Gimme a few mins.
I've reuploaded the source tarball for those having the ADL_SDK problem. Please try again. |
|
|
|
I tested 2.7.6 on my rig over the day and I'd like to thank you for great work. Latest release (compared to 2.5.0) also fixed some issues with occasional re-uploading of shares, so now I finally see ultra-low stales with 2.7.6 + Stratum proxy.
Thanks. There are an awful lot of changes from 2.5.0 to 2.7.6. |
|
|
|
I was just thinking since the servers are set to give out 12 getworks per minute now vs. 8 getworks per minute on the old dynamic test server, that I would see a new "U" value of 12 (instead of the 16 that I'm seeing).
I don't think the estimation from the server will be perfect to give you exactly the right number of getworks. It will only ever be approximate within a range of timeframes. If you run cgminer 2.7.6 it will show you what diff shares you are submitting now. |
|
|
|
|
Show Posts
