Page 2 of 5

Re: No more W/Us for Radeon 4xxx

PostPosted: Sat Apr 21, 2012 9:52 pm
by AndrewDixon
Thanks for the replies, at least that explains it :(

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 1:33 am
by somata
Just wanted to add that I have a Radeon HD 4850 and am still getting units, so as Nathan mentioned there are units out there. However, yesterday I waited about 30 minutes for a WU before giving up. Even today, while I've gotten several WUs, there's typically a five to twenty minute gap between them while it's waiting for an assignment. So the WU supply does seem to be dwindling.

Also I wanted to correct verlyol; V7 does indeed support core 11 (GPU2). I've been using the V7 client with GPU2 for a few months with no trouble other than some of the stats missing in the work unit pane (such as credit and likewise PPD, though TPF does show up).

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 6:52 am
by GreyWhiskers
Being the "quant" that I am, I wanted to shed a little light on just what I'm seeing on my HD4670 AGP running on v7.1.52.

I went into the HFM stats it's been gathering since it became v7 compatible. Under v7, the normal mode is that you download the next WU when you reach a certain percentage of completion of the last. Default is 99%, although I have all my computers set to 100%- that works excellently for all but the "bigadv"s, which I'm no longer getting on my i7 2600k.

anyway, when the Work server 171.64.65.103 started slowing down, there could be a considerable wait until it had new WUs to hand out. Most of the time, the Assignment Server would indicate "empty work server assignment", or other error codes. the v7 client will patiently wait to retry getting an assingment, but it sets an increasingly long timeout till it tries next, making some of the turn around times very long.

Anyway, here's the data set.

re: format. First, the download time, next the completion time, then the "turnaround time". The latter is the difference between the download time and the previous WU's completion time, finally the PRCG.

If the work server can send a new WU as soon as asked, the turnaround time is negative, the download occurs between the time the previous WU reaches 100% and the time when processing is complete (packaging of the results for upload to Stanford and cleaning up after the processing takes that time).

But, when the work server can't hand out enough WUs fast enough, the numbers become positive - the next WU isn't downloaded until some time after completion of the previous. the longest in this dataset is over 15 hours waiting.

Since the norm lately is lengthy waits, multiple hours, I've thrown in a Uniprocessor core to take advantage of the dead time. This is kind of a crap shoot. Back when the 511 point WUs were all that were being served on the Core11, the GPU processing dovetailed very well with the Uniprocessor processing on my old single-core Pentium 4/HT 3.2 GHz system. Unfortunately, the 4724 WUs don't tolerate a uniprocessor sharing the computer very well - ppd drops off dramatically; the 4756s tolerate the interference much better.


Code: Select all
  DownloadDateTime   CompletionDateTime Turnaround time       PRCG
                                          hh:mm:ss.00
  3/23/2012 03:47     3/23/2012 11:18                   4724, 9, 178, 0
  3/23/2012 11:17     3/23/2012 16:13     -00:00.56.07  4724, 9, 346, 3
  3/23/2012 16:13     3/23/2012 20:23     -00:00.21.12  4724, 6, 466, 73
  3/23/2012 20:22     3/24/2012 00:33     -00:00.27.84  4724, 8, 603, 0
  3/24/2012 00:32     3/24/2012 05:00     -00:00.26.01  4724, 10, 615, 0
  3/24/2012 04:59     3/24/2012 09:17     -00:00.25.02  4724, 10, 253, 5
  3/24/2012 09:17     3/24/2012 13:39     -00:00.19.35   4724, 9, 72, 2
  3/24/2012 13:38     3/24/2012 17:50     -00:00.26.03  4724, 8, 370, 2
  3/24/2012 17:50     3/25/2012 03:05     -00:00.25.05  4756, 2, 980, 67
  3/25/2012 03:04     3/25/2012 07:33     -00:00.32.25  4724, 11, 403, 12
  3/25/2012 07:32     3/25/2012 11:52     -00:00.31.89  4724, 9, 481, 3
  3/25/2012 11:52     3/25/2012 16:12     -00:00.23.49   4724, 9, 81, 6
  3/25/2012 16:12     3/25/2012 20:25     -00:00.10.52  4724, 11, 127, 5
  3/25/2012 20:25     3/26/2012 00:34     -00:00.09.44  4724, 10, 546, 4
  3/26/2012 00:33     3/26/2012 04:55     -00:00.29.91  4724, 11, 493, 6
  3/26/2012 04:55     3/26/2012 09:09     -00:00.33.45   4724, 9, 10, 2
  3/26/2012 09:09     3/26/2012 13:26     -00:00.17.96  4724, 9, 561, 5
  3/26/2012 13:26     3/26/2012 17:46     -00:00.33.16  4724, 9, 742, 0
  3/26/2012 17:45     3/26/2012 22:07     -00:00.10.36  4724, 10, 737, 0
  3/26/2012 22:06     3/27/2012 02:31     -00:00.24.60  4724, 10, 755, 0
  3/27/2012 02:30     3/27/2012 06:43     -00:00.20.86  4724, 10, 349, 3
  3/27/2012 06:42     3/27/2012 11:01     -00:00.33.49  4724, 8, 218, 12
  3/27/2012 11:00     3/27/2012 15:17     -00:00.27.94  4724, 10, 486, 8
  3/27/2012 15:16     3/27/2012 19:49     -00:00.30.03   4724, 8, 25, 8
  3/27/2012 19:49     3/28/2012 00:04     -00:00.18.96  4724, 4, 665, 66
  3/28/2012 00:04     3/28/2012 04:35     -00:00.29.20  4724, 11, 862, 1
  3/28/2012 04:35     3/28/2012 08:57     -00:00.25.86  4724, 9, 279, 14
  3/28/2012 08:57     3/28/2012 13:08     -00:00.34.56  4724, 10, 209, 7
  3/28/2012 13:07     3/28/2012 17:23     -00:00.36.74  4724, 11, 834, 3
  3/28/2012 17:22     3/28/2012 21:52     -00:00.45.60  4724, 10, 815, 4
  3/28/2012 21:52     3/29/2012 02:11     -00:00.17.67  4724, 11, 602, 5
  3/29/2012 02:11     3/29/2012 06:23     -00:00.37.86  4724, 9, 242, 15
  3/29/2012 06:34     3/29/2012 14:28     +00:10:36.56  4724, 11, 555, 6
  3/29/2012 14:27     3/30/2012 02:20     -00:00.42.28  4724, 9, 485, 9
  3/30/2012 02:19     3/30/2012 11:14     -00:00.40.77  4724, 8, 384, 11
  3/30/2012 11:13     3/30/2012 19:37     -00:00.33.49  4724, 8, 688, 7
  3/30/2012 19:36     3/31/2012 04:18     -00:00.30.41  4724, 10, 228, 18
  3/31/2012 04:17     3/31/2012 12:59     -00:00.15.25  4724, 9, 393, 14
  3/31/2012 12:59     3/31/2012 21:30     -00:00.05.60  4724, 8, 731, 4
  3/31/2012 21:29      4/1/2012 05:29     -00:00.10.94  4724, 8, 785, 3
   4/1/2012 05:29      4/1/2012 14:27     -00:00.22.52  4724, 8, 643, 12
   4/1/2012 14:26      4/1/2012 18:55     -00:00.13.03  4724, 11, 968, 1
   4/1/2012 18:55      4/1/2012 23:08     -00:00.24.45  4724, 11, 596, 8
   4/1/2012 23:08      4/2/2012 03:22     -00:00.20.55  4724, 10, 930, 2
   4/2/2012 03:22      4/2/2012 07:31     -00:00.18.77  4724, 8, 756, 5
   4/2/2012 07:31      4/2/2012 11:39     -00:00.14.64  4724, 10, 676, 14
   4/2/2012 11:39      4/2/2012 15:52     -00:00.19.86  4724, 10, 904, 3
   4/2/2012 15:51      4/2/2012 19:58     -00:00.32.50  4724, 11, 776, 10
   4/2/2012 19:58      4/3/2012 00:06     -00:00.12.28  4724, 9, 816, 6
   4/3/2012 00:06      4/3/2012 04:13     -00:00.30.33  4724, 8, 15, 24
   4/3/2012 04:12      4/3/2012 08:27     -00:00.21.48  4724, 8, 755, 17
   4/3/2012 08:26      4/3/2012 12:36     -00:00.25.16  4724, 8, 273, 20
   4/3/2012 12:36      4/3/2012 16:58     -00:00.07.97  4724, 9, 496, 28
   4/3/2012 16:58      4/3/2012 21:05     -00:00.23.56  4724, 10, 642, 10
   4/3/2012 21:04      4/4/2012 01:14     -00:00.21.39  4724, 8, 835, 5
   4/4/2012 01:14      4/4/2012 05:25     -00:00.31.38  4724, 9, 954, 11
   4/4/2012 05:25      4/4/2012 09:35     -00:00.36.00  4724, 10, 992, 4
   4/4/2012 09:35      4/4/2012 13:49     -00:00.27.17  4724, 10, 574, 7
   4/4/2012 13:48      4/4/2012 19:09     -00:00.15.34  4724, 8, 903, 5
   4/4/2012 19:12      4/4/2012 23:25     +00:02:07.08  4724, 8, 483, 20
   4/4/2012 23:27      4/5/2012 03:38     +00:02:20.81  4724, 8, 598, 18
   4/5/2012 03:37      4/5/2012 07:55     -00:00.31.86  4724, 8, 369, 19
   4/5/2012 07:56      4/5/2012 12:12     +00:00:27.50  4724, 9, 233, 41
   4/5/2012 12:12      4/5/2012 16:24     -00:00.18.48  4724, 9, 529, 23
   4/5/2012 16:23      4/5/2012 20:51     -00:00.23.73  4724, 9, 962, 5
   4/5/2012 20:51      4/6/2012 03:05     -00:00.19.62  4724, 11, 238, 40
   4/6/2012 03:04      4/6/2012 12:38     -00:00.35.86  4756, 8, 999, 60
   4/6/2012 12:38      4/6/2012 16:59     -00:00.43.65  4724, 10, 564, 24
   4/6/2012 16:59      4/6/2012 21:55     -00:00.40.07  4724, 11, 721, 20
   4/6/2012 21:54      4/7/2012 07:30     -00:01.02.42  4756, 3, 999, 51
   4/7/2012 07:32      4/7/2012 12:04     +00:02:19.69  4724, 10, 773, 22
   4/7/2012 12:03      4/7/2012 16:23     -00:00.35.60  4724, 8, 316, 17
   4/7/2012 16:28      4/7/2012 21:21     +00:04:43.04  4724, 8, 624, 15
   4/7/2012 21:23      4/8/2012 02:04     +00:01:51.68  4724, 11, 728, 16
   4/8/2012 02:03      4/8/2012 11:37     -00:00.44.98  4756, 6, 983, 84
   4/8/2012 11:36      4/8/2012 16:08     -00:00.35.04  4724, 11, 732, 24
   4/8/2012 16:07      4/9/2012 01:31     -00:00.39.34  4756, 0, 971, 121
   4/9/2012 01:33      4/9/2012 06:24     +00:02:11.62  4724, 11, 853, 17
   4/9/2012 06:43      4/9/2012 11:21     +00:18:33.74  4724, 10, 640, 19
   4/9/2012 11:22      4/9/2012 15:41     +00:01:47.77  4724, 8, 94, 30
   4/9/2012 15:57     4/10/2012 01:26     +00:15:49.74  4756, 2, 996, 84
  4/10/2012 01:28     4/10/2012 11:08     +00:01:45.40  4756, 2, 838, 127
  4/10/2012 11:07     4/10/2012 15:29     -00:00.23.49  4724, 8, 755, 39
  4/10/2012 15:28     4/10/2012 19:58     -00:00.34.27  4724, 10, 781, 32
  4/10/2012 19:58     4/11/2012 00:26     -00:00.30.46  4724, 9, 484, 31
  4/11/2012 00:28     4/11/2012 05:15     +00:01:49.46  4724, 10, 564, 40
  4/11/2012 05:19     4/11/2012 10:09     +00:04:38.12  4724, 9, 758, 25
  4/11/2012 10:08     4/11/2012 14:55     -00:00.52.41  4724, 9, 485, 54
  4/11/2012 14:57     4/11/2012 19:44     +00:02:01.04  4724, 8, 504, 59
  4/11/2012 19:59     4/12/2012 05:23     +00:15:39.69  4756, 9, 874, 106
  4/12/2012 05:23     4/12/2012 09:40     +00:00:09.85  4724, 10, 75, 54
  4/12/2012 09:45     4/12/2012 14:36     +00:04:32.94  4724, 8, 706, 29
  4/12/2012 14:35     4/12/2012 19:09     -00:00.33.48  4724, 10, 645, 52
  4/12/2012 19:17     4/12/2012 23:46     +00:08:46.12  4724, 8, 679, 37
  4/13/2012 00:31     4/13/2012 10:31     +00:44:30.92  4756, 0, 999, 87
  4/13/2012 10:47     4/13/2012 19:06     +00:15:30.16  4724, 8, 664, 33
  4/13/2012 19:08     4/14/2012 00:09     +00:01:51.96  4724, 9, 972, 48
  4/14/2012 00:09     4/14/2012 05:28     -00:00.30.45  4724, 9, 766, 39
  4/14/2012 07:29     4/14/2012 11:53     +02:00:52.80  4724, 10, 536, 42
  4/14/2012 12:09     4/14/2012 16:47     +00:15:38.96  4724, 10, 346, 59
  4/14/2012 17:02     4/15/2012 02:56     +00:15:37.49  4756, 5, 746, 132
  4/15/2012 07:53     4/15/2012 12:50     +04:57:19.75  4724, 9, 936, 45
  4/15/2012 12:50     4/15/2012 17:28     -00:00.27.11  4724, 9, 879, 23
  4/15/2012 20:45     4/16/2012 01:30     +03:16:40.69  4724, 10, 481, 38
  4/16/2012 01:30     4/16/2012 06:16     -00:00.23.22  4724, 9, 884, 50
  4/16/2012 06:15     4/16/2012 11:01     -00:00.46.80  4724, 11, 413, 68
  4/17/2012 02:17     4/17/2012 10:25     +15:16:05.50  4724, 9, 906, 57
  4/17/2012 10:24     4/17/2012 17:51     -00:00.45.08  4724, 11, 690, 61
  4/17/2012 20:45     4/18/2012 03:09     +02:54:38.77  4724, 10, 481, 47
  4/18/2012 04:22     4/18/2012 08:43     +01:13:32.94  4724, 11, 491, 76
  4/18/2012 09:57     4/18/2012 14:18     +01:13:57.77  4724, 9, 758, 55
  4/18/2012 15:39     4/18/2012 19:55     +01:21:08.55  4724, 11, 900, 58
  4/19/2012 07:27     4/19/2012 11:46     +11:32:43.30  4724, 8, 929, 68
  4/19/2012 23:25     4/20/2012 09:33     +11:39:29.17  4756, 5, 765, 100
  4/20/2012 10:17     4/20/2012 21:02     +00:44:21.81  4724, 10, 481, 59
  4/21/2012 01:44     4/21/2012 17:38     +04:42:25.71  4756, 0, 955, 118


Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 12:45 pm
by AndrewDixon
Thanks for the insight Grey.

Going forward are there any viable low cost ATI/AMD options that produce similar performance but possibly supported for a while longer?
What sort of PPD do the newer low end cards e.g. HD5450 / HD6450 / HD6570 produce?

For reference my 4830 was producing around 1400PPD.

Reason for not wanting to spend much is just to make use of my 2nd GPU that sits basically idle simply to connect 2 extra monitors (Have a Radeon 4890 as my main card which is plenty for what bit of gaming I do).

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 1:11 pm
by De Berghut
AndrewDixon, that's a question I'd also like to ask myself. My budget is extremely low at this time, and for the very little gaming I do I don't need no top of the bill GPU anymore.. Sure it would be nice, but as the budget is limiting itself... ;) I know it doesn't sound logical on the money-point, but I'm considering buying an Nvidia Card again. On the game-side, I've been more than happy with my card. But I often experienced quite some issues and problems, which I absolutely never had with Nvidia.. And I often get the feeling Ati isn't so interested any more about it's historical (once exclusive even) engagement with FAH... :?

Does anybody know if there will be some kind of "official" announcement of the final support-stop on GPU2? :( It's gonna hurt to put my Radeon HD 4870 asleep :cry: So much wasted resources... :e(

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 1:37 pm
by 7im
Pande Group always makes announcements when ending support for hardware.

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 4:41 pm
by p24601
I had the same problem with my HD4850, I was getting only one WU a day (with v7). I was very lucky on Friday I got two :shock:
I was lucky enough to have been able to get a new card, HD6570 on Saturday and I've not looked back.

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 4:46 pm
by p24601
AndrewDixon wrote:Thanks for the insight Grey.

Going forward are there any viable low cost ATI/AMD options that produce similar performance but possibly supported for a while longer?
What sort of PPD do the newer low end cards e.g. HD5450 / HD6450 / HD6570 produce?

For reference my 4830 was producing around 1400PPD.


Hi, I've just bought a HD6570 and I'm getting around 3261PPD

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 7:14 pm
by exit151
I was just going to put up a post about my GPU (abn ATI 4890) being stuck in 'download' as it's status, where the log shows
"******************************** Date: 22/04/12 ********************************
16:27:45:WARNING:WU00:FS00:Failed to get assignment from 'assign-GPU.stanford.edu:80': Empty work server assignment
16:27:45:WARNING:WU00:FS00:Failed to get assignment from 'assign-GPU.stanford.edu:8080': Empty work server assignment
16:27:45:ERROR:WU00:FS00:Exception: Could not get an assignment"

over an over.. Guess this is why? I'm curious why this decision was made.. The 4890 is hardly "old and antiquated" and delivers suburb stats/processing power, which is great for gaming and conversely been great for folding.. Due to a driver issue we were "screwed" if we wanted to fold a few months ago dating back for several months, which has now been fixed, and now suddenly no more work for it? That really stinks. I cannot possibly imaging the processor (4890 and/or whatever else has been cut) becoming useless and not worth the time to set jobs up for. I gotta say, days after my post about why I fold, this is very displeasing. It'd be one thing if this 'was' an old card, but crap, it's what, 2 generations back, and beats performance-wise many of the generation it was replaced by!

Any fellow ati 4xxx owners feeling this pain as well??

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 8:12 pm
by Nathan_P
The problem is not the power of the card, its the GPGPU language that it uses. Nvidia have used cuda since the orginal 8800 gtx and apart from tweaks (especally from 1.0 to 1.1) it has remained a constant and compatatble language

ATI/AMD on the other hand used an language called Brook for 2xxx, 3xxx exclusively and also in part on 4xxx gpu's. This was difficult to program for so they stopped using it and switched to openCL. This forced PG to write the new ATI core,core 16 for openCL which is supported on all 5xxx and above cards. Unfortunately 4xxx gpus only have limited support for openCL so have unfortunately been relegated to obsolesence when the current core 11 work is done

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 8:41 pm
by bruce
7im wrote:Pande Group always makes announcements when ending support for hardware.


I've forgotten, but I think they already made that announcement. If so, we should find it and reference it directly.

Re: No more W/Us for Radeon 4xxx

PostPosted: Sun Apr 22, 2012 11:37 pm
by Zagen30
bruce wrote:I've forgotten, but I think they already made that announcement. If so, we should find it and reference it directly.


From March 15th, 2011:
http://folding.typepad.com/news/2011/03/fah-support-for-ati-gpus.html

For the short term, we will support both the older (Brook coded) ATI core 11, and the newer (OpenCL coded) ATI core 16, but we cannot support the Brook-based core for much longer. We plan to support ATI core 11 until September 1, 2011 (hopefully longer, but we want to get this on donors' radars).

Re: Only 1 WU per day for Radeon 4xxx?

PostPosted: Sun Apr 22, 2012 11:59 pm
by GreyWhiskers
The notice came over a year ago - see Vijay's blog - of the impending deplenishment of folding support with these series of old ATI cards. As I see it, we are almost 8 months past the announced 1 Sept date, and they haven't pulled the plug yet, although there seem to be more users attempting to get Core 11 WUs than the available supply. There must have been a good reason for new projects 4724 and 4756 to have been rolled out in the past several months - good science being done.

It would be good to know what the PL strategy is - keep limping along on the 4724 and 4756 projects until all the science those projects need is exhausted and then once an for all, pull the plug; or are there going to be more new projects on the horizon?

From: A blog all about Folding@home, from its Director, Prof. Vijay Pande
March 31, 2011
Core 16 for ATI released; also note on NVIDIA GPU support for older boards

In a previous post, I mentioned our plans for supporting ATI GPUs. I'm happy to announce that with the release of v7 into open beta, we have now released ATI core 16 GPU3 WUs into advanced methods for Folding@home. Please see Dr. Lin's forum post for details.

We are doing a gradual rollout of these WUs, for a couple of reasons. First, while this has gone through beta testing, it is still pretty new, so we want to be cautious with its rollout. Second, I would like to add some more servers to help the load. These servers are being delivered today and will likley be on line in a week or two. Depending on the load on the ATI core 16 Work server, we may wait for a full release until the new hardware is on line.

Just a reminder from the previous post that there are some limitations in our ability to support ATI boards. In particular, we're limited to those which support OpenCL 1.1 in order to get any sort of reasonable performance out of the hardware. This means that only 5xxx ATI GPU boards or later will work with the new core 16 (the series 3xxx are not supported and the series 4xxx does not have sufficient OpenCL support). However, we will keep core 11 support going hopefully at least until Sept 1, 2011 in order to support the older hardware.
Also, we are looking at support of older NVIDIA GPUs as well. In particular, there have been reported issues with some 8xxx series boards. We may have to limit their support as well due to analogous hardware compatibility issues.

We're excited to be able to release OpenCL support for ATI GPUs. However, I want to stress that it is still pretty early days for this code, but we're excited about where this is going.


Posted at 10:57 AM in code development, FAH achievements | Permalink

Re: Only 1 WU per day for Radeon 4xxx?

PostPosted: Mon Apr 23, 2012 1:43 am
by 7im
Like I said, they always make an announcement. ;)

Re: No more W/Us for Radeon 4xxx

PostPosted: Mon Apr 23, 2012 11:37 pm
by De Berghut
exit151 wrote: Due to a driver issue we were "screwed" if we wanted to fold a few months ago dating back for several months, which has now been fixed, and now suddenly no more work for it? That really stinks. I cannot possibly imaging the processor (4890 and/or whatever else has been cut) becoming useless and not worth the time to set jobs up for.
Any fellow ati 4xxx owners feeling this pain as well??


Even if I find it hard to "digest", I can see why they're doing it.. I've been hoping they'd find a way to using the slight present openCL on for example my HD4870, but I guess it's been faint hope ;)

Anyway, I read they fixed the driver issue? I've updated just a few weeks ago, 3 at most, still had the old problem though... EUE's all the way.. Rolled back: all ran smoothly again.. :shock: