Message boards :
Number crunching :
I am back (who? II)
Message board moderation
Author | Message |
---|---|
ML1 Send message Joined: 25 Nov 01 Posts: 20437 Credit: 7,508,002 RAC: 20 |
Ok, I was rather surprised at the "I am back" thread getting closed... OK, various hype from Francois. But then, getting optimisations to work is 'hard'. Good skepticism from most other people. And a few good discussion points. And thanks to the support that Intel give or have given to the s@h effort. Meanwhile, it's all up to who? Myself, it will be interesting to see and to discuss what the numbers say. Who? Regards, Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
Crunch3r Send message Joined: 15 Apr 99 Posts: 1546 Credit: 3,438,823 RAC: 0 |
If Francois decides to post here, it's his decision and be welcome. ANY bashing towards AMD or INTEL is NOT appropriate in this thread here, so keep that in mind ;) Anything else will go to "Recycle Bin" ;) Thanks. Join BOINC United now! |
ML1 Send message Joined: 25 Nov 01 Posts: 20437 Credit: 7,508,002 RAC: 20 |
OK, so is this perhaps all just a case of high passions drifting into fervor...? There is also a slight 'disconnect' between the dispassionate engineering side of the discussion and then the more emotional comments from a "lay person's" viewpoint. There has been many a 'confusion of understanding' due to a clash of culture and/or one word of miscommunication... (Was that burp a dire insult to your grandmother's cooking or a sign of great appreciation of the meal?...) I'm sure who? can play the forums as he wishes. It will still be very interesting for what he comes up with, and how. Happy crunchin', Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
The Gas Giant Send message Joined: 22 Nov 01 Posts: 1904 Credit: 2,646,654 RAC: 0 |
I think it will be very interesting seeing the actual comparisons between Penryn and Barcelona. In the mean time who? is the only "man on the inside" that makes himself available for comment and we need to take his comments as such. For me it's all a mute point in any case as I normally skip a cpu generation or two as replacing computers at each new generation is just too expensive, but reading about their capabilites and where the technology is heading is more than a little interesting. I encourage who? to continue to make comment. I just wish there was someone from within the AMD camp to give feedback on their status, but I hazard a guess they do not want to give Intel a heads up on how good their product is until it is released. With some luck Intel will pre-emptively release Penryn in large amounts prior to the release of Barcelona and then release Nehalem shortly after.....good times are a coming! Live long and BOINC. Paul (S@H1 8888) And proud of it! |
ML1 Send message Joined: 25 Nov 01 Posts: 20437 Credit: 7,508,002 RAC: 20 |
I think it will be very interesting seeing the actual comparisons between Penryn and Barcelona. In the mean time who? is the only "man on the inside" that makes himself available Very much so... And I think it would be very unwise for an AMD rep to announce himself/herself on this forum... Meanwhile, a little Penryn hype backed up by real numbers is interesting. ... With some luck Intel will pre-emptively release Penryn in large amounts prior to the release of Barcelona and then release Nehalem shortly after.....good times are a coming! And there's the catch... It's all a question of "timing"... So far, there's only "ES" parts circulating. And there are rumours that AMD have functionally correct Barcelona parts but that their clock speeds are 'unexciting' at the moment due to unfavourable critical paths in the logic. The real fun begins when we can actually get them into motherboards for ourselves. I may well put together a PC components based 8 core NUMA beast myself to test out a few clever inter-CPU comms tricks to see how well it works. I'm looking to put some servers together so I may as well get a few extra bits to 'have a play'. The decider may well be whether I can more easily get a group of AMD chips or a group of Intel chips to usefully talk to each other... Mmmmmmm... Regardless of the raw numbers, it would be good if "who?" could divulge what ideas/algorithm he's been chasing for the last year or so... ;-) It's all useful and good computer science/engineering. Happy crunchin', Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
Brock Send message Joined: 19 Dec 06 Posts: 201 Credit: 774,488 RAC: 0 |
The real fun begins when we can actually get them into motherboards for ourselves. How about a bit of premature advice? When you get your hands on a Barcelona or Agena quad-core, make sure that the Node Interleave (NI) is set properly for what the O/S environment supports. That can have a big performance impact on CPU memory cross-talk in a NUMA environment. It has to do with process thread assignment to nodes. A node is basically a CPU socket (not a core). The idea is to limit memory access to 0 hops (on the same memory controller) instead of having a lot of 1 hop (1 controller away) or 2 hop (2 controller away) memory accesses. The memory hops go over the conherent HyperTransport connections that connect AMD Opteron processor sockets together. You can have up to 1 memory hop operations on a dual-socket motherboard and 2 memory hop operations on a quad-socket motherboard. Brock |
John Clark Send message Joined: 29 Sep 99 Posts: 16515 Credit: 4,418,829 RAC: 0 |
I think it will be very interesting seeing the actual comparisons between Penryn and Barcelona. In the mean time who? is the only "man on the inside" that makes himself available for comment and we need to take his comments as such. For me it's all a mute point in any case as I normally skip a cpu generation or two as replacing computers at each new generation is just too expensive, but reading about their capabilites and where the technology is heading is more than a little interesting. I encourage who? to continue to make comment. I just wish there was someone from within the AMD camp to give feedback on their status, but I hazard a guess they do not want to give Intel a heads up on how good their product is until it is released. With some luck Intel will pre-emptively release Penryn in large amounts prior to the release of Barcelona and then release Nehalem shortly after.....good times are a coming! Like you TGG, I replace rigs after several generations have gone bye. So, fortunately I have 3 to replace over the next 3 - 4 years, and next year I hope to replace my first - a single core P3 Coppermine with a Penryn. The following year my dual P3 will see a Nehalem unit replace it. That then leaves my dual Xeon Prestonia to replace, which is 4 years old, with the generation around then. It's good to be back amongst friends and colleagues |
Ghery S. Pettit Send message Joined: 7 Nov 99 Posts: 325 Credit: 28,109,066 RAC: 82 |
Well, you might want to check Who?'s RAC. It has taken a rather serious jump in the past day. I suspect we may be seeing the results of his efforts very soon. |
kittyman Send message Joined: 9 Jul 00 Posts: 51469 Credit: 1,018,363,574 RAC: 1,004 |
Well, you might want to check Who?'s RAC. It has taken a rather serious jump in the past day. I suspect we may be seeing the results of his efforts very soon. Go who?!!! Got a link to the computer ID? "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
Well, you might want to check Who?'s RAC. It has taken a rather serious jump in the past day. I suspect we may be seeing the results of his efforts very soon. With the way the Top Computers list is behaving, you'll probably have to wait a couple more days for the V8 to reach the top 20. It's no longer 3122418, I presume who? wants to get enough work done to properly demonstrate its performance. Joe |
Kim Vater Send message Joined: 27 May 99 Posts: 227 Credit: 22,743,307 RAC: 0 |
Who?'s computer are now hidden?? ;) Kiva Greetings from Norway Crunch3er & AK-V8 Inside |
Philadelphia Send message Joined: 12 Feb 07 Posts: 1590 Credit: 399,688 RAC: 0 |
Who?'s computer are now hidden?? ;) I haven't checked, but if that true, it s#@ks. What's he hiding??? |
Astro Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
There might be hints as to what he's trying in the "result ID", and perhaps it's not time to reveal that to the world just yet. Or, perhaps all the work is validating badly and he doesn't want that known. I don't know. |
Francois Piednoel Send message Joined: 14 Jun 00 Posts: 898 Credit: 5,969,361 RAC: 0 |
Who?'s computer are now hidden?? ;) when you get a computer on the top without owner, you ll know what happen :) my opti is yet OFF :) ... and yes, i am validating. no need to expose this. will take few days. who? |
Francois Piednoel Send message Joined: 14 Jun 00 Posts: 898 Credit: 5,969,361 RAC: 0 |
Who?'s computer are now hidden?? ;) I merge with last code available too. |
B-Roy Send message Joined: 4 May 03 Posts: 220 Credit: 260,955 RAC: 1 |
Who?'s computer are now hidden?? ;) I think that who? will show his work when he is ready to do so. We should rather cheer him up a bit, cause any new optimization will help this project to be more effective and yield more credit for users who are dedicated to the project. In the end, who? also has a full-time job, so it's great he actually spends so much time with this. but I can understand everybody waiting without patience, since who? is drumming the marketing machine quite loudly. Let's see what we get and how much you can optimize this code further. |
Keith T. Send message Joined: 23 Aug 99 Posts: 962 Credit: 537,293 RAC: 9 |
Who?'s computer are now hidden?? ;) They're not hidden at Rosetta: http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=192677 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=341221 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=366022 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=377010 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=384409 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=417462 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=418931 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=424658 http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=540850 Sir Arthur C Clarke 1917-2008 |
B-Roy Send message Joined: 4 May 03 Posts: 220 Credit: 260,955 RAC: 1 |
Who?'s computer are now hidden?? ;) I guess that won't help much if we are curious about his new Seti application. |
ML1 Send message Joined: 25 Nov 01 Posts: 20437 Credit: 7,508,002 RAC: 20 |
... and yes, i am validating. no need to expose this. Hey! The validation period is just as interesting as the final optimised run period. You need to give what the optimisation speedup is... Are your optimisations due to better compilation for the Penryn? Or are there algorithm optimisations? Standard Penryn or overclocked? Good luck, Regards, Martin See new freedom: Mageia Linux Take a look for yourself: Linux Format The Future is what We all make IT (GPLv3) |
The Gas Giant Send message Joined: 22 Nov 01 Posts: 1904 Credit: 2,646,654 RAC: 0 |
He can keep it as quiet as he likes.....but, I am sure he will give us all the gory details once it comes to a conclusion! |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.