The new #1 host.. AMD? :-D - EDIT: yes! :-)
Message boards :
Number crunching :
The new #1 host.. AMD? :-D - EDIT: yes! :-)
Message board moderation
| Author | Message |
|---|---|
 Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 
|
A nice 32 Core rig.. Quad-Core AMD Opteron(tm) Processor 8354 now at #73.. but maybe soon #1? 8 x Quad-Core !! Nice.. nice.. :-D With my QX6700 @ 4 x 3.14 GHz I had in well times a RAC of ~ 6,400. ..AK v8.0 inside.. :-D A ~ 74 Credit WU in ~ 3,600 seconds The 8354 @ 32 x 2.21 GHz a ~ 74 Credit WU in ~ 13,000 seconds Both with AK v8.0 ! :-D ~ 3.5 times longer but 8 times more Cores.. 6,400 / 3.61 x 8 = 14,182.82.. So speculation if the AMD is running 24/7 with 100 % SETI@home a RAC of: 14,182.82 ?? Huhhh.. then we have a new #1 rig.. WITH AMD !! :-D 
ID: 785924 ·  |
nick  Send message Joined: 22 Jul 05 Posts: 284 Credit: 3,902,174 RAC: 0 
|
that thing has been working less than a month....and has more than 1/2 the credit i have gotten in 3 years...WOW! i want one...but i would not want to pay for power for it.... lets wait and see if it beats me to 270000 credits here.  
ID: 785938 · |
Pilot  Send message Joined: 18 May 99 Posts: 534 Credit: 5,475,482 RAC: 0 |
A nice 32 Core rig.. Looks like both AMD and Intel are using the same path to keeping Moores law intact. Just kep doubling the number of CPU's on the core. I have seen some info on 32, 64, and 128 CPU's on single chip core from intel, but I will believe it when it happens. IMHO both are up aginst a rock and a hard place to feed these beasties over the current bus design. AMD has good design but now I don't think they can really feed 32 CPU at higher speeds. Intel has increased bus speed but will run into a rock wall as they add more CPU to cores. I think next big step will be 1 bus for each CPU per core. Place yer bets ladies and gentelmen. I will volunteer to hold all wagers for a small fee;) When we finally figure it all out, all the rules will change and we can start all over again. ID: 786188 · |
OzzFan     Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28
 |
I think next big step will be 1 bus for each CPU per core. With HyperTransport or QuickPath Interconect, I'd say 1 bus per 4 cores would be more sufficient without being wasteful of bandwidth. Maybe even 1 bus per 8 cores depending on future revisions of each technology's ability to deliver data at high speeds. Of course, doing this decreases the efficiency of inter-core communications and cache coherency. It would be ideal if a special HT or QPI was designed especially for inter-core communications too (or something faster than HT or QPI). ID: 786388 · |
|
Pilot Send message Joined: 18 May 99 Posts: 534 Credit: 5,475,482 RAC: 0 |
I think next big step will be 1 bus for each CPU per core. You are most likely right as to the number of bus needed per core or CPU. I just get frustrated with disk drives and other slow devices at times. Multiple pipes do allow easier overlap of operations at the same time. When we finally figure it all out, all the rules will change and we can start all over again. ID: 786391 · |
|
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5
|
A nice 32 Core rig.. #61 now.. :-D
ID: 786588 · |
|
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5
|
A nice 32 Core rig.. Quick update, because of changing the side of the overview.. ;-) Now #58.. RAC 7,452.80 :-D
ID: 786683 · |
|
Dennis Send message Joined: 26 Jun 07 Posts: 153 Credit: 15,826,319 RAC: 0
|
A nice 32 Core rig.. It is shameful, IMHO, it is taking this 32 core AMD rig to do just slightly more than an 8 core MacDaddy or about the same as a pair of quad Q9450's are doing... BUT if this is being done as a test bed of some kind I can't think of anything more demanding or a better test environment on a processor on a 24/7 basis that SETI/BOINC. ID: 786733 · |
|
nick Send message Joined: 22 Jul 05 Posts: 284 Credit: 3,902,174 RAC: 0
|
um...ya...i lost badly in that race....hes almost at 170000, i am at 153000.....yay! for amd for beating 2 low end core 2s. lol anyways, does anyone know what type of motherboard it might have?
ID: 786803 · |
Ivailo Bonev  Send message Joined: 26 Jun 00 Posts: 247 Credit: 35,864,461 RAC: 2 
|
ID: 786820 · |
|
Pilot Send message Joined: 18 May 99 Posts: 534 Credit: 5,475,482 RAC: 0 |
A nice 32 Core rig.. Ha ha ha just like life eh?? It doesn't matter how many men you put on the job, it still takes 9 months to get the baby;) Im not sure where we are going with all the multi core/cpu platforms. It seems to me that the more eggs you put in one basket, the more likely you are to have a diaster. When we finally figure it all out, all the rules will change and we can start all over again. ID: 786830 · |
|
H Elzinga Send message Joined: 20 Aug 99 Posts: 125 Credit: 8,277,116 RAC: 0 
|
does anyone know what type of motherboard it might have? I have to take a gues but here it goes. I only know of motherboards able to handle 4 processor (4*4=16) which would leave us with a gap of 16 cores. What options are left wil be found in the server market. Something like this A5808 or this VX50. These are based on a Tyan motherboard (see post by Ivailo) which can cary 4 aditional processors on a daughter board.This is unsuitable for home use (unless you buy a compatible encloser which has enough room for this board and 2 psu 
ID: 786845 · |
|
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5
|
A nice 32 Core rig.. Now #51.. RAC 7,664.17 :-D Yes of course it's not very well that the AMDs are not so fast here at SETI@home.. :-( I was before SETI@home an AMD fan, because I want to support the small manufacturer. But to support on the best way SETI@home was to buy an Intel. But the reasons to my knowledge that the AMDs are so slowly are among other things because of the smaller L2-Cache and that the optimizer software support better the Intel CPUs.
ID: 787086 · |
john deneer  Send message Joined: 16 Nov 06 Posts: 331 Credit: 20,996,606 RAC: 0
|
Something like this A5808 or this VX50. Why would that be 'unsuitable for home use'? Especially in home use there is no need to build stuff into an enclosure, a table with enough space will suffice ;-) Regards, John. BTW and very off topic: you wouldn't happen to be the Hans Elzinga that studied chemistry in Utrecht in the 1975-198x period? ID: 787101 · |
|
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5
|
ID: 787684 · |
RottenMutt  Send message Joined: 15 Mar 01 Posts: 1011 Credit: 230,314,058 RAC: 0
|
may hit a rac of 20K  
ID: 787800 · |
Mumps [MM]  Send message Joined: 11 Feb 08 Posts: 4454 Credit: 100,893,853 RAC: 30
|
may hit a rac of 20K Well, a current rough guesstimate based on AndyK's page would be about 12,475. ID: 787804 · |
kittyman  Send message Joined: 9 Jul 00 Posts: 51469 Credit: 1,018,363,574 RAC: 1,004
 |
may hit a rac of 20K Not too bad, considering the Frozen Penny according to AndyK would be over 10k on just 4 cores....smokin'.....LOL. "Freedom is just Chaos, with better lighting." Alan Dean Foster 
ID: 787808 · |
|
nick Send message Joined: 22 Jul 05 Posts: 284 Credit: 3,902,174 RAC: 0
|
may hit a rac of 20K never seen this site you speak of, do you have a link? thanks
ID: 787853 · |
|
kittyman Send message Joined: 9 Jul 00 Posts: 51469 Credit: 1,018,363,574 RAC: 1,004
|
Too trashed, ask me in tbe morning.......or somebody else will chime in........... "Freedom is just Chaos, with better lighting." Alan Dean Foster
ID: 787857 · |
©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.