http://img514.imageshack.us/img514/6676/id2ep6.jpgQuote:
Originally Posted by hargen
Printable View
http://img514.imageshack.us/img514/6676/id2ep6.jpgQuote:
Originally Posted by hargen
I dunno if im misinterpreting the numbers but those results seem slow to meQuote:
Originally Posted by s7e9h3n
I think it does not that bad.
This is from a Athlon 64 @ 3500 MHz.
Especially the sse load operation seems to be much much faster than the Athlon 64's. 2 loads/cycle vs 0.5 loads/cycle....
That's FANTASTIC. This is what I get on a Sempron 3000+ :D :Quote:
Originally Posted by xPliziT
Those scores seem accurate for the K10, as it can do 2 128-bit loads per cycle.
The Core 2 can only do 1 load per cycle I believe.
Looks great to me, here's one from a Opteron 185.
s7e9h3n, what about
http://slil.ru/24945619
http://slil.ru/24945621
to test?
only problem with this bench is that its non coherent. by a high margin, not just variation between runs.
here is ss from my 5200+@3.17Ghz, everytime it gives way another results lol...
i would preffer to see PCmark05 CPU tests results
Dunno what these benches prove as they're single threaded tests ;)Quote:
Originally Posted by MAS
http://img530.imageshack.us/img530/6938/mashn3.jpg
whether they single or multythreaded - that doesn't matter
they show the raw power of a core
here is K8 2GHz result
http://img208.imageshack.us/img208/3...6987987ag9.jpg
I was referring more to the bench that told me my Barcelona performed like a k8 venice @ ddr400 =P.Quote:
Originally Posted by MAS
here is an old coolaler's ss
you can compare results
conroe is obviously a little faster
the program generates and verifies electronic signature on elliptic curves, this is pure INT test
http://img520.imageshack.us/img520/4773/image001fg0.gif
Tisk tisk, where are PC MARK cpu scores?
Where are 3D MARK cpu scores?
Where are NUCLEARUS runs?
Where are WINRAR benches?
Where is CINEBENCH scores?
This SSE drivel is all fine and well for the truely boring people, but REAL-WORLD users wanna see REAL-WORLD numbers.
C'mon.....lets run a benchmark that people know and can relate to.
In order:Quote:
Originally Posted by SOLDNER-MOFO64
Will run when I install 32bit XP
Thought I gave those already?
Ok...will put it + Winrar in the queue gonna finish this WCG bench first..
Thought I ran that as well?
TBH, I don't find anything interesting in a lot of these benches myself, but some people don't have their own Barcelona setups to bench for themselves. All I can do is try to accomodate as many requests as I can and when I finally get my L1n64 to work with these chips, I'll run what I want to run :yepp:
I tested it with Core 1 and Core 2 for comparison and the results are always the same.Quote:
Originally Posted by s7e9h3n
Core 1:
http://img267.imageshack.us/img267/8...ore1lh9.th.png
Core 2:
http://img524.imageshack.us/img524/4...ore2vu9.th.png
Do you mean not the same? :confused: Those results you posted are both different to one another, first one is worse than a P4 2.8GHz, second one is very slightly better.Quote:
Originally Posted by pumero
These short tests seem to have a large inaccuracy when being run, since they're that quickly over - just look at the deviation.
OK...finally found BA rig for comparison....take a look here:Quote:
Originally Posted by KTE
http://www.google.com/translate?u=ht...&hl=en&ie=UTF8
Here's a summary of his results using a pair of BA 2347's on a Tyan board:
First bench mark results (syntetisch):
SiSoft Sandra IIX (tested under servers 2003 hr. 32-Bit)
and/or two CCUs put
49.800/24,726 Dhrystone ALU (MIPS)
48.612/24,363 Whetstone FPU (MFLOPS)
7.166/4,282 MB/s memory range
58.839/27,382 MB/s Cache/Speicherbandbreite
143.637/71,667 iit/s Multimedia Integer x4 aEMMX/aSSE
184.910/94,299 fit/s Multimedia floating point x4 iSSE2
Cinebench 10 (1-Core/8-Core, tested under Longhorn 32-Bit/64-Bit)
32-Bit CB 10064
64-Bit CB 1813/12383 points factor: 6.83
Cinebench 9,5 (1-Core/8-Core, tested under Longhorn 64-Bit)
64-Bit CB 271/1519 points factor: 5.61
SuperPi
superpi 1M = 43s
3DMark2006 (tested tied up with ATI X1900XT 256MB, over PCI-E x8)
5463 points
From those results, I guess it's fair to conclude that the BA's ARE different than the B!'s, BUT not in the way that everyone (besides me :p: ) has been ASSUMING to this point. I don't know about the Sandra Benches, but if you compare BOTH tsuehpsyde AND dave_graham's B1 Cinebench results from earlier in this thread with this guy's BA's, you'll see that the B1's results OBLITERATED the BA's. And for good measure, dave_graham even beat the BA's 1m score with his B1's. I know it's not exactly a mountain of evidence, but it's something ;)
Upon further review, it looks like tsuehpsyde's B1's beat the BA's across the board in Sandra (with the exception of the memory bandwidth bench which is quite erratic.)
Also note that the guy is running the same mobo as Dave....;)
I'll have to look through the other results. Two things right off the top of my head though, I couldn't tell if he got it to run in dual chanel and He was running the 2347s not the 2350s you guys are running. He's also running the card in a X8 slot and my experience you'll lose score.
I'm the only one running 2350's here....everyone else has been running 2347's - that's why I wasn't using my results to compare to ;) And my PCIE runs @ x8 as well.....Quote:
Originally Posted by PhilDoc
EDIT: And IF he did happen to accomplish running in Dual Channel for those benchmarks, that would make it even more sad:p:
You're right, my mistake. Just assumed everyone got the 2350s for review. The results are sad.
s7e9h3n that's awesome mate! :)
I have some evidence against him now to pass on (because he's been arguing like a child oppositely). Its enough to show the point quite clearly although I did see the guy talking about waiting for a new Tyan BIOS further on in the thread with sad faces, which in sign language would tell me something is not too perfect. :D
If all went perfect with the system then that shows B1 as representative of BA, simply put.
KTE he put words properly... Basically he tested Core Duo (32bit Intel CPU) and then he tested Core 2 Duo (64bit Intel CPU) :p:Quote:
Originally Posted by KTE
Besides this test is very consistent, at least for me! Maybe it has to do something with OS or BIOS that some of you are getting big result variations?:confused:
@pumero Thanks for your input! :up:
@s7e9h3n I'm really grateful for your work! If I would live in US then you've got a beer on me (or if you ever come to UK :))! Still waiting for some OC!
PS. Can you try RealStorm benchmark?
No, I believe the 667mhz is indicating the speed of his DDR2. DCT0 and DCT1 are memory channels 0 and 1 on each of the 2 onboard chipsets. My Bios version?Quote:
Originally Posted by MR_SmartAss
http://img206.imageshack.us/img206/3033/cpuzmobonw2.jpg
But note that the mobo I'm running isn't currently available anywhere. In fact, it hasn't yet officially been announced. My board is a Tyan s3993 built on the yet-to-be-released Serverworks/Broadcom HT2100 Chipset....;)
Dunno if this is correct, but here's Winrar: (note this bench is single threaded)
http://img186.imageshack.us/img186/5550/winrarqm3.jpg
and here's that Realstorm bench - which, btw, isn't correctly identifying my hardware....
http://img186.imageshack.us/img186/2061/tssc1.jpg
just thought i'd toss my 2 cents in here. there were two bios' floating around for the Barcelona before release...one offered significantly more performance than the other but it wasn't production. the actual pre-production bios was more consistent in performance (i.e. perf wasn't wildly across the spectrum) but was "slower" in measurable performance.
yes, BIOS' can have significant performance impact on CPUs.
in any case, i'm still trying to work through some day job work but if anyone's in Chicago on this coming wednesday night, let me know...I'll be in town.
dave
Now, that wasn't exactly what I was trying to prove as there ARE differences between BA's and B1's. But, like I've been saying, these differences are relegated to specific situations and seemingly don't appear in the tests that we currently are comparing....;)Quote:
Originally Posted by KTE
Agreed about the Bios comment, but may I ask which bios you were running on your s3992?Quote:
Originally Posted by dave_graham
Use the latest Beta version of WinRAR S7e9h3n ( 3.71 Beta 1 ), it's multithreaded and the fastest out there [ most optimized for newage cpus ].
I apologize if I missed it, because I have been trying to follow this thread as closely as I can. Have there been any results using WCG work units on either a 2P 2347 or 2350 - with or w/o oc? Even boinc benchmarks would be helpful. Just trying to get an idea for what a Phenom 4x4 rig might be able to do.
Yep, ran Bionc earlier today. While these aren't spectacular, expect the numbers to get better as support for this mobo appears...Quote:
Originally Posted by twilyth
http://img529.imageshack.us/img529/7123/wcg3tr5.jpg
EDIT: just noticed you're an XS cruncher....check the WCG lounge for a few more SS's ;)
THx for the tip....Better, but still didn't load the cpu's up to 100%. Usage hovered around 60% max during the bench:Quote:
Originally Posted by BenchZowner
http://img444.imageshack.us/img444/2105/winrar2qz9.jpg
That was, low had expected a bit more form a 4x4
Got 1330KB/s with X2 3600+ 'Brisbane'@2835MHz - Ram 472 mhz 4-4-4-8-11 T2.
Yeah i know bad BIOS, mobo, ram and so on, but still had expected over 2000 from a 4x4
Does your cpu usage reach 100%? I have a feeling the bench is optimized for 4 or less cores.....Quote:
Originally Posted by nullface
s7e9h3n could you please run this benchmark?
http://www.7-zip.org/
Download the beta and post some screenshots.
Since you can adjust the thread count you can see how well the system scales.
Press the stop button after the total rating appears.
Here are my results
So once you can overclock, you should be getting about 1800 mflops per core at 2.2ghz? That's assuming 10% over stock - but I'm sure you'll be pushing the envelope.Quote:
Originally Posted by s7e9h3n
I don't know if I did the math right but if an oc'ed 6600 gets a boinc benchmark of 2400 mflops and 13,000 pts per day then a 2P 2350 getting 1800 mflops should do about 20,000 pts per day ((2400 * 4)/13000 = (1800 * 8)/x). I think that sounds about right.
Will stay tuned.
http://img130.imageshack.us/img130/9997/7zipmw8.jpgQuote:
Originally Posted by xPliziT
set "bench.exe" process priority to realtime in order to get true result - some processes obviously bother the benchQuote:
Originally Posted by MR_SmartAss
as to mentioned 25 per cent this is wrong
Conroe has only ~ 10 per cent advantage over K8 in INT instructions and up to 100 per cent in SSE instructions - in sum that gives Conroe AVERAGE advantage of 20 per cent over K8 - it's a well-known fact
Andrew
The winRAR result is pretty good, and the realstorm result is bad - 1.87 GHz Conroe 2M: 6.88 fps, 3.01 GHz 10.9 fps. The matrix multiply results are really good and the 2 loads/cycle is fantastic, but something just isn't adding up. WinRAR performance would be severely affected by registered 667 Mhz single channel RAM.Quote:
Originally Posted by s7e9h3n
Could it be the fact that realstorm detects 0mb of system memory and doesn't recognize any l3 cache on the cpu? :p:Quote:
Originally Posted by JVguest
Edit: The SS doesn't show the L3 info, but trust me on this
Awesome job s7e9h3n. I would appreciate it if you can make a run with the Folding@Home SMP client below. :D If you get around to benchmarking it, be sure to post the Project #. Just let it run for 2% and we should be good.
Thanks
Quote:
Originally Posted by JVguest
Oh yeah...I actually had installed it, but never got around to running the bench since it required my login info for windows. I have that feature disabled on my OS, so I have to setup a new user account and password before I can get it to run.....I'll get it done....;)Quote:
Originally Posted by SP33DFR34K
Awesome! But just to warn you, if you didn't install the client recently, you are going to have to get the new copy. The beta for the last one expired Oct 1.Quote:
Originally Posted by s7e9h3n
no worries as I ended up uninstalling the old one anyway as it kept running in the background. I must have installed @ least 20-30 progs in the last 24 hours so I can run these benches for u guys...hopefully, they'll all be just as easy to uninstall :)Quote:
Originally Posted by SP33DFR34K
s7, so, from your understanding:
BA are almost the same as B1?
and,
Still the same bios problems with ccNUMA and Dual-Channel?
Also, for curiosity: still nothing from the L1N front, correct?
Thanks a lot for all new info! :D
Its gotta be a piss take to have the hardware but no bios option to overclock :(
how hot are these quads at load at their default voltage ?
thanks in advance
Yep, pretty much so...I'd think you'd be pretty hard pressed to find some benchmarks which exploit the differences between the two. I don't doubt that there were changes made to the cpu between the steppings, but the problems which were fixed probably were rather insignificant in terms of most applications. Come to think of it, I recall a few times where I've had early stepping AMD's which never made it to production and they performed just as well/if not better than their retail counterparts (at least in the realm of the benchmarks which i ran).Quote:
Originally Posted by JohannesRS
Yep, still getting no support from AMD or Tyan. I really wish they would get on the ball as this board is being wasted in its crippled state....:rolleyes:
http://img442.imageshack.us/img442/1386/ht3vq4.jpg
Also, Realstorm failed to detect SSE2 and SSE3.Quote:
Originally Posted by s7e9h3n
~30C after 5hours @ 100% load :cool:Quote:
Originally Posted by Slay0r
Thanks for the information, s7. Really a shame these problems with ccNUMA and Dual-Channel
What's your ambient temperature??? And processor clock??:eek: :shocked:Quote:
Originally Posted by s7e9h3n
@Dave: Dave, could get something running in any mobo that included all HT links and the energy split?
Thanks a lot you both! Your information is absolutelly "unique" for all us! :) :up:
Dave got 1573KB/s with the 2347s back on page 7 for me. ;)Quote:
Originally Posted by s7e9h3n
Quote:
Originally Posted by KTE
winrar should be good on these chips. Irrc, Anand's review showed the 2350 to be equivalent to a Xeon at 2.7G. It was the Barcelona's best showing.
Hey Stephen, can you run kribibench? Here's from my T7400ES laptop with 2gb ram @ 533mhz 4-4-4-12. Will toss up a woodcrest score in a bit.
Kribibench: http://www.adeptdevelopment.com/
edit
Woodcrest 5120's @ 2709mhz with 4x1GB FB-DIMM's @ 386mhz 5-5-5-15, i5000x chipset snoopfilter enabled. (@2.33ghz scores were 2.88fps using a remote viewing application to bench)
What about some S&M running? It should rape those poor opterons to their max. ;)
(Not a bench, more like a CPU toaster)
http://www.benchmarkhq.ru/fclick/fclick.php?fid=301
s7e9h3n, since you don't appear to be a Linux guy, what about installing OpenSSL for Windows (light version: only 1 MB) ? Get it from [1]. Then assuming you install it in the default location (C:\OpenSSL), open a command prompt and run this (if you get an error "MSVCR71.DLL could not be found", install Microsoft .NET Framework 1.1 [2] and copy C:\winnt\microsoft.net\framework\v1.1.4322\msvcr71 .dll to C:\openssl\bin\):
"openssl speed" should take about 10 min to complete (single threaded by default, the Linux version can be run in multi-threaded mode), it benchmarks:Code:C:\> cd openssl\bin
C:\> openssl speed
md2, md4, md5, hmac(md5), sha1, sha256, sha512, rmd160, rc4, des, triple des, aes-128, aes-192, aes-256, idea, rc2, blowfish, cast, 512-bit rsa, 1024-bit rsa, 2048-bit rsa, 4096-bit rsa, 512-bit dsa, 1024-bit dsa, 2048-bit dsa.
All these algorithms are 100% ALU code (except on Linux, where sha512 uses SSE2), some are coded in C, others are in assembly. They all scale pretty much linearly with the clock frequency. They should all easily fit in the Barcelona L2 cache so they should not be influenced by the memory subsystem at all. The only downside of the Windows version is that it is compiled as a 32-bit executable. Some algorithms would see a significant gain by running in 64-bit mode: RSA and DSA would be about 3 times faster, RC4 30% faster, MD5 15% faster, etc. Anyway see below for a table of results I obtain on a old Pentium M 1.2 GHz: the fastest hash algorithm is MD4 (240 MB/s) and the fastest symmetric cryptographic algorithm is RC4 (173 MB/s).
Others should run this benchmark on Intel Core-based CPUs so we can compare against Barcelona...
[1] http://www.slproweb.com/download/Win...ght-0_9_8e.exe
[2] http://www.microsoft.com/downloads/d...7-034D1E7CF3A3
- ZCode:--- openssl speed results for Pentium M 1.2 GHz on 32-bit Windows ---
OpenSSL 0.9.8e 23 Feb 2007
built on: Wed Feb 28 01:35:20 2007
options:bn(64,32) md2(int) rc4(idx,int) des(idx,cisc,4,long) aes(partial) idea(int) blowfish(idx)
compiler: cl /MD /Ox /O2 /Ob2 /W3 /WX /Gs0 /GF /Gy /nologo -DOPENSSL_SYSNAME_WIN32 -DWIN32_LEAN_AND_MEAN -DL_ENDIAN -DDSO_WIN32 -D_CRT_SECURE_NO_DEPRECATE -D_CRT_NONSTDC_NO_DEPRECATE -DBN_ASM -DMD5_ASM -DSHA1_ASM -DRMD160_ASM -DOPENSSL_USE_APPLINK -I. /Fdout32dll -DOPENSSL_NO_CAMELLIA -DOPENSSL_NO_RC5 -DOPENSSL_NO_MDC2 -DOPENSSL_NO_KRB5 -DOPENSSL_NO_DYNAMIC_ENGINE
available timing options: TIMEB HZ=1000
timing function used: ftime
The 'numbers' are in 1000s of bytes per second processed.
type 16 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
md2 858.35k 1797.09k 2502.85k 2744.40k 2863.25k
mdc2 0.00 0.00 0.00 0.00 0.00
md4 9161.87k 31980.97k 92768.68k 175310.51k 240031.58k
md5 8152.99k 28237.34k 79240.60k 145100.25k 190056.26k
hmac(md5) 10780.19k 35380.04k 91967.75k 145100.25k 190056.26k
sha1 7774.79k 24811.94k 60322.57k 94813.31k 113321.28k
rmd160 6563.34k 19407.97k 42649.42k 61353.87k 70065.63k
rc4 139810.13k 164676.25k 171511.10k 173175.23k 172640.63k
des cbc 24185.12k 24973.53k 25206.15k 25934.79k 25747.72k
des ede3 9060.93k 9176.90k 9272.75k 9249.70k 9224.45k
idea cbc 17098.67k 18045.84k 18291.77k 18448.67k 18418.29k
rc2 cbc 8794.01k 9978.72k 10044.43k 10053.46k 10053.46k
rc5-32/12 cbc 0.00 0.00 0.00 0.00 0.00
blowfish cbc 40223.49k 42441.73k 43118.01k 43656.56k 43464.29k
cast cbc 38347.92k 40672.04k 41455.93k 41620.48k 41775.94k
aes-128 cbc 35091.44k 36897.33k 36033.54k 37282.70k 37282.70k
aes-192 cbc 31034.44k 31956.60k 32634.15k 32640.50k 32634.15k
aes-256 cbc 27535.23k 28787.26k 28941.20k 28941.20k 29021.30k
camellia-128 cbc 0.00 0.00 0.00 0.00 0.00
camellia-192 cbc 0.00 0.00 0.00 0.00 0.00
camellia-256 cbc 0.00 0.00 0.00 0.00 0.00
sha256 5540.33k 13459.46k 24542.45k 31236.67k 33872.84k
sha512 1843.35k 7146.39k 11545.61k 16405.63k 18504.62k
sign verify sign/s verify/s
rsa 512 bits 0.001189s 0.000100s 840.7 9986.3
rsa 1024 bits 0.005844s 0.000300s 171.1 3329.8
rsa 2048 bits 0.035111s 0.001049s 28.5 953.4
rsa 4096 bits 0.234300s 0.003362s 4.3 297.5
sign verify sign/s verify/s
dsa 512 bits 0.000940s 0.001159s 1063.5 862.5
dsa 1024 bits 0.002772s 0.003363s 360.7 297.3
dsa 2048 bits 0.009327s 0.010798s 107.2 92.6
--- end ---
Saw this from David Kanter over at RWT. Barc non _rate SPEC benches.
http://realworldtech.com/forums/inde...83478&roomid=2
Good god, those #'s make the Barcelona look exceedingly bad.Quote:
Originally Posted by Periander6
Quote:
Originally Posted by Periander6
The link doesn't seem to be working
worked for me.
wish it didn't. it'll make me have nightmares tonight. :P
Works fine for me. If you want you can try just going to the main RWT site then going to the forums. Not hard to find Kanter's post there about it.Quote:
Originally Posted by PhilDoc
Beware guys, in these non-rate SPEC results, Intel actually make use of the Intel compiler auto-parallelization feature when AMD does not (AFAIK), meaning that you are comparing apples to oranges (code automatically multi-threaded by the compiler on Intel vs. mono-threaded code on AMD).
Besides, these non-rate AMD numbers have been available for about a month at [2].
It looks like Intel is working very hard at producing the best SPEC results possible. A few weeks ago the best SPECfp2006 score for a dual Xeon X5365 3.0 GHz was 16.9 [3], and just now they publish a score of 21.4 with the same CPUs [1] (which is the real news brought by this realworldtech.com post). Look at the peak score of the subtest 436.cactusADM for example: it jumped from 20.1 to 95.9 just by using the -parallel compilation flag (check it for yourself in [1] and [3]).
What Intel is doing does not violate the SPEC rules (they explicitely allow auto-parallelization), but it makes any sort of direct comparison of the performance of Opteron vs. Xeon impossible.
[1] http://www.spec.org/cpu2006/results/...903-01960.html
[2] ftp://ftp.software.ibm.com/eserver/b...cpu_091007.pdf
[3] http://www.spec.org/cpu2006/results/...820-01861.html
- Z
I search the non-paralle single-thread test of intel. The search results of non-auto-paralle shock me:
SPECfp2006(base/peak)
AMD 1.9GHz - 10.7/11.2
Intel 1.86GHz - 12.3/12.5 (test data: 2007.6.25)
http://www.spec.org/cpu2006/results/...625-01332.html
Intel 3GHz - 18.7/19.3 (test data: 2007.7.23)
http://www.spec.org/cpu2006/results/...723-01537.html
To surprise, the non-paralle result of SPECint2006 is much faster than the paralle result.
SPECint2006(base/peak)
Intel 3GHz 20.2/22.6 (test data: 2007.7.23)
http://www.spec.org/cpu2006/results/...723-01539.html
Which does one rule the single-thread test?
hargen: you link to SPEC results of mono-processor machines (Xeon UP 3040 and C2D E6850). They use faster unbuffered DDR2 RAM, which give them an advantage compared to the 2P results (Opteron 23xx and Xeon 53xx) we have been discussing so far (most of the SPECfp2006 tests are very memory-intensive).
- Z
SPECfp2006 rates tests are very memory-intensive. But SPECfp2006 tests of non-paralle are single-thread floating-point tests,and more depend on the core raw power. Xeon 3040 1.86G uses ecc ddr2 ram, but opteron uses ecc ddr2 ram too. The difference isn't that large - 2% or less?
I have to agree that "AMD will need to increase the frequency substantially to keep competitive with Intel on the desktop", because single-thread or two-thread performance will rule the desktop in future 20 months.
Depends on the application. For the average desktop user, Quad Core is overkill. Browsing the web, spreadsheets and word-processing don't require anything other than duel core.Quote:
Originally Posted by hargen
For a gamer, quad core will rule the roost. All of the latest game coming out in the next 20 months will take advantage of multiple cores.
I've no doubt a 2.6-3.0ghz Phenom will be competitive with equally priced Intels but I very much doubt we will see a AMD take the performance crown for a long time to come.
Speaking of the Phenom. Are any of these chips actual Duel cores or are they all Quads but with two core disabled?
Do not use games to determine the performance of the desktop. game tests are favour of intel: games use too many branch, load/store-dependency-chain and so on. Only 10-20% of the instructoins in games are floating-point instructions. game tests are biased.Quote:
Originally Posted by Motiv
Oh WOW, I love the Sig LOL!:rofl:Quote:
Originally Posted by MR_SmartAss
@hargen
Yes, games should be used since many of us here are Gamers and give our computers a good workout. When they brought up games, they were talking about Desktops, not Servers.
I'm waiting on Quad Core 45nm and that's why I didn't go with the G0 C2Q.
Quote:
Originally Posted by Periander6
Thanks, finally got it to work, don't know what the problem was. Yep, pretty sad, still haven't seen anything to turn me off or on either way. Besides, I still have a few months to decide.
What concerns me more is the inability to get the systems to run correctly
My post was in response to you saying single threaded will be the way forward to the next 20 months.Quote:
Originally Posted by hargen
This is not correct. Quad core will be the way forward from now on. Most applications other than basic Office apps, are making use of multicore. The more cores, the more it will use.
Dual core will be the norm for low spec machines but expect quad core to be mainstream.
I looked at the SPEC scores in details, and I don't understand where you see that the "non-paralle result of SPECint2006 is much faster than the parallel". Here is a SPECint2006 result with auto-parallelization for a 3 GHz Xeon [1]: 20.9/24.3. This is better than your non-parallel 20.2/22.6.Quote:
Originally Posted by hargen
[1] http://www.spec.org/cpu2006/results/...916-02014.html
That's running the -parallel flag for C Benchmarks base optimization as well. ;)Quote:
Originally Posted by zpdixon
Yeah, I am not sure but I think that vendors are free to do what they want in the base and peak tests.Quote:
Originally Posted by KTE
- Z
I posted this in the news thread as well.
Quote:
Here's some quick K8 results @2GHz specint base=9.77 rate=10.8
specfp base=10.4 rate=10.9
This is way too close. Either there is something wrong with the K10 IBM system or K10 is going to be a huge disappointment for everybody except the HPC crowd. I would wait for at least a couple more results to get published. Just look at the problems stephen has been having with the various Bioses.
http://www.spec.org/cpu2006/results/...828-01902.html
http://www.spec.org/cpu2006/results/...828-01900.html
Um why we looking at single threaded benches on a Native Quad Core,Im confused.The power of the platform is that all the cores can talk to each other very quickly.The only thing I dont understand is the SSE should be twice as fast as K8,even on a single thread,One thing I know for sure is on my Quad FX ,setting memory speed to less that 800,even with lower latency timings slows it down,also setting the bios interleave all(disabling NUMA)kills it.It seemed S7 has that problem.Comparing an Intel with slower memory timing proves nothing because the platforms are totally different.In my tests nothing can use a dual socket AMD system as well as SUSE linux.Windows XP x64 sometimes takes 20 minutes to figure out its not using local memory and then move the thread so it does.Vista wasnt much better.Off node hits kill the AMD 2P .The Desktop Quad AMD will be using HT3 and have alot more memory bandwidth than Barcelona, I also heard that at a certain clock speed it has some kinda power band that kicks in that amazing,since the person who said that was one of the first to provide bench marks .Im gonna guess thats right.Barcelona had to plug into an old socket and run and it does but not as fast as it will in a new socket.I know this Barcelona scales better than lineir that has never before happened in the history of processors,example 1 processor scores 50,2 processors score 110,history says 2 processors are about 30% faster than 1 or 65Quote:
Originally Posted by BrowncoatGR
Yeah they did. AMD officially stated that you can expect "up to 15% performance gain" going from 2GHz to 2.5GHz. Those were server market releases and based on upcoming SPEC CPU2006 figures though. Nothing of desktop performance was indicated at all. And they're planning to launch a 2.6GHz Opteron 23xx model by end '07 too.
Quote:
Originally Posted by KTE
If you look at the spec web site, AMD has put up some number for the 2360SE and I belive the 2357.
Just to make it easier for people. Thanks for the heads up Phil...Quote:
Originally Posted by PhilDoc
Spec FP Rate (8 cores)
SpecFP Rate (16 Cores)
SpecIntRate (8 cores)
SpecIntRate (16 cores)
2360SE it's 18%-22% faster then 2347... not bad, perfect scaling would be 25%.Quote:
Originally Posted by PhilDoc
At 8 cores it performs better thana X5355 in rate benches. I'm curious about ST performance though.
@Smartass unless you can prove that Quad xeons communicate faster through the FSB than K10 does internally i say you are making that up.
Quote:
Originally Posted by Motiv
Np, was looking at them thought people would like to see them. Thanks for posting the nice links
Enrol yourself into our NEW CAMPAIGN !!! :eek:
It's Codenamed WFAL (pronounced Waffel) campaign and you'll love it :up:
What do we do in our campaign? :D
We WFAL !!! Or WAIT FOR ACTUAL LAUNCH !!!!
Don't compare C2D and Barcelona, wait for launch of desktop parts this Nov/Dec. :mad:
Perkam
Quote:
Originally Posted by perkam
Well, its called curiosity and there are a few server people on here.
Btw, you could use a spell checker.
I noticed this when it was published.Quote:
Originally Posted by MR_SmartAss
So, AMD K8 dual core 2.2GHz communicates between each core taking 53.5ns and the K10 quad core 2.0GHz communicates taking 76ns... according to those tests. And thus Netburst->Core 2 halved the previous latency while K8->K10 added ~50% to the latency...hmmm. Anyone else reproduce these tests?
Here was another Barcelona vs Harpertown short review: http://www.besttechreview.com/content.php?IndexID=750
Quote:
Originally Posted by KTE
This is an exact copy of an Anand article
http://www.anandtech.com/IT/showdoc.aspx?i=3099
As far as communication between cores. By definition of the program, memory systems would come into play and who knows how accurate the program is too begin with. Finally, I'm starting to feel like Perkam, this is just getting to be a little obsessive. Time to wait for the Phenom launch.
LOL! I've found 6 other articles in Eastern sites that were copied off ATs Barcelona reviews, word for word. :DQuote:
Originally Posted by PhilDoc
You know, I've felt fed up ever since September 10th, but I still want to see real numbers for fact. And I was only lurking on Opteron 2300 going by this thread. Phenom is yet far off IMO, and that only changes for me when someone gets one in their hand. ;)Quote:
As far as communication between cores. By definition of the program, memory systems would come into play and who knows how accurate the program is too begin with. Finally, I'm starting to feel like Perkam, this is just getting to be a little obsessive. Time to wait for the Phenom launch.
I've got one question tonight...
Ok, Barcellona uses HT1.0.
Phenom uses HT3.0
Phenom FX uses HT3.0
What about FASN8???????????? :D :D :D
I was meaning no offense, just stating my own frustration :up:Quote:
Originally Posted by KTE
Non taken bro, don't worry. :)
this thread is the source of my frustration for the past couple of weeks... why can't asus get a working bios...
Another question:
if I would build today a dual 23xx based opteron, with two processors, around 4Gb of RAM (that would possibly increase with time) and two HDs... What sort of PSU should I use? "How much power" it should have?
Thanks a lot.
Zippy 700, PCP&C 750?
Without a vid card, a class 400W is more than enough as you'll see in reviews for peak VAC consumption.