OK OK, but we don't have a gaming board to test them on now. Do we? Maybe if one only used two cores?
Printable View
OK OK, but we don't have a gaming board to test them on now. Do we? Maybe if one only used two cores?
It's nice to see the great Kyosen excited about a cpu (especially an AMD one) :up: but I must beg to differ here. I have a feeling CPU-Z will show a "1" in the "stepping" for a revision BA cpu ;) ......of course I could be wrong here :p:Quote:
Originally Posted by kyosen
This is without a doubt the most interesting topic on XS right now. I won't post anything more as you guys are out of my league. ;) Carry on...
(btw Dave Graham, I've added you to my MSN.)
O right, now things are much more clear!Quote:
Originally Posted by metro.cl
Just like Hexus, THG, and some other sites you're hurt 'cos you wasn't "awarded" with samples, demo stuff, or presence at the launch, and now you've started to grown small, but fast growing hatred towards the AMD and their product… and you even find yourself smart enough to discuss with Dave, although you know he has whole world of experience with Barcelona, and you didn't even touch the HSF!!
And now you're telling us that you have your own little vendetta plan – they don't giving me samples, then I'll leak! And being such a big leaker does you really expect that AMD will give you any support?? Get a grip amigo, your attitude is completely wrong!
And please, don’t mention Kubicki… we all know how lame he went in discussion with Charley, and we all know how blatantly he was making up FUD regarding the Phenom system photos… he's gone in the wind of subjectivity and hurt ego ('cos he wasn't invited)… don't go that route!:shakes:
Lol AMD attitude is completely wrong.Quote:
Originally Posted by Nedjo
And as far as i know the website is the only one in chile with this award.
http://www.chilehardware.com/images/embajadoresAMD.jpg
I still get samples but have to return them, just to let you know i went to AMD headquarters this year invited by them to the HD 2900 launch. I always said i would definetly stay here and get samples that stay so we can compare to NVIDIAs when both hardware is available.
My current VGA is a HD 2900 XT 1GB so you might have the wrong idea.
Now i'm not a fanboy i used to love my Opty 148 and did some awsome overclocking with it, my current notebook is based in AMD cpu + chipset. I just like when people say the truth.
Ask Dave if he knows Kris maybe you dont know much about things really.
About K10 CPUs of course i would never get samples down as is the normal thing from AMD, we are always working to improve this, but AMD doesnt care about SA.
Last time i checked leaked info was something nice to other users, maybe you just dont have to look at info when i leak it if you hate that much my attitude.
You are really out of line with this post my friend., Go and chill out a little.Quote:
Originally Posted by Nedjo
yea, I'm a mod defending a XIP, but that isn't the issue. Your tone is.
So much good information in this thread. But so much thread crapping as well. It seems people get blinded by their ego's over hardware here and who's "right". It's like Dave and S7 have said..the hardware still isn't really ready. There are a few bugs to work out (as with every CPU launch) and I'm not surprised at all. For AMD not having the vast resources of Intel to pull a Native Quad off and get it going right away is impressive in itself. Along with their acquisition of ATI, AMD is doing the right thing. All that matters is that their orders have shipped to the Enterprise customers which is where the $$$$ is at.
Talking about perfomance scale. I've made some numbers of perf scale with techreport review (I've only jumped synthetic and useless crap Sandra). Here we go:
http://img487.imageshack.us/img487/9...ortbarcfo6.jpg
Well, it seems Barcelona scales clearly better than Clovertown (18.61% of 25% or 74.44% of 100% in Barcelona side; 18.23% of 28.76% or 63.40% of 100% in Clovertown side). So, it scales about 11% more than Clovertown.
Now the evident question: Is it because Barcelona 2.5ghz is B2 and Barcelona 2.0ghz is B1? or, is it simply because Barcelona scales better than Clovertown? here's the question. Someone?
I don't see anything insulting in my post and don't see the reason for you to defend metro 'cos I wasn't attacking him, or using any "banana" words!Quote:
Originally Posted by Movieman
It's clear as a day from his recent posts and factless arguing with Dave, that metro is… well, let's say disappointed with Barcelona, but hasn't really got any real supporting argument for that state of spirit. And with that attitude he's engaging confrontation with Dave who has all the hard facts about CPU, and it's performance…
And finally metro is writing to Dave that he's sold to AMD (BTW why didn't you reacted when metro accused Dave that he's "sold to AMD"?) . So I've asked myself why he's acting so frustrated and finally he's giving us an answer:
" Read this i believe you know Kris and also i leak because as i've said before is the only thing i can do since we don't get the hardware we want even thought we are like Top 5 hardware review website in Spanish."
And now he's giving us evidence that he was really recognized by Advanced Micro TWO years ago, he has loved his Opteron, it's allegiance to AMD is true – he has true rarity 2900XT 1GB etc.
Unfortunately AMD doesn't recognize that any more, although he's trying to improve that, unfortunately without results, 'cos "AMD doesn't care about SA"…
Back to the beginning: you don't need to defend metro, 'cos I was giving him friendly advice not to go with the route of leaking, 'cos that will not improve his relationship with AMD, quite contrary! And believe me I know what I'm writing about!!
Also he should stop paying attention what Kubicki is writing about 'cos that boy screws up badly his reputation, and not any call upon rebellion against NDA will not improve things! Reputation is the only weapon that IT jurno has and when it's lost once it's lost for good. And if metro is making living from his site he'll be smarter than Kubicki.
So once again, metro, you should try to work out things with AMD if you want to be recognized by them, leaking will not get you anywhere, and in the long run will diminish aggregated clicks on account of leakage.
Concerning the K10, you haven't see but fragment of info in those PDF's that you've got, about upcoming stuff
Cheers m8
Wish this thread was more about truth in numbers than soap operas......losing interest fast.
You got a short memory. Go back and read your post. The tone is sarcastic and out of line. I also douby any post to an individual that ends with a :down: is meant to be positive or helpful.Quote:
Originally Posted by Nedjo
Yup, the thread badly needs some cleaning :shakes:
Been trying hard to separate the information from the poo in these threads, but it is becoming harder and harder. Thanks to the guys asking and answering questions; and to the others, please keep the shoveling to a minimum.
My memory is just fine! As I wrote: I haven't used single word to insult him. I've wrote what's obvious: metro has a problem in communication with AMD and has reflection his aggravation with a company (or local office or representative) on to the product, regardless of the fact that he doesn't have clear picture of the product (unlike Dave who he calls "sold to AMD)Quote:
Originally Posted by Movieman
I've miss clicked smiley! Edited post for appropriate one…Quote:
I also douby any post to an individual that ends with a :down: is meant to be positive or helpful.
Chill out guys (all). You're on the internet and its pointless getting serious. No ones gonna harm you here, well at least not physically.... I skip over watching TV you know. :D
PetNorth: I'll check your numbers out after some system building I gotta do right now. Looks good at a glance.:D
The problem, and yes, I believe it is a problem folks, that this chip is server market. Server market, what the various chip PR teams play up, hype and fight over is what counts most: usually the SPEC CPU 2006 benchmarks first, power second and simultaneously, price.
In the same respect, when AMD says its Opteron 2.5GHz scales 15% better than a 2GHz, I assure you they mean for the market segment and applications its to be sold in: server. I really reckon AMD went for 2S, 4S, 8S kills with this silicon features such as HT 3.0/DCA2.0, massive bandwidth power and low TDP. Because that's where most of the profit is and day by day, even more than 4-way supercomputing and general computing power requirements are increasing in demand. In these sections, power efficiency and price is key with decent performance, believe me.
Look at the successful uptakings already: Hewlett-Packard, IBM, Sun, Dell, Appro, Egenera, Gateway, Rackable Systems, Supermicro, and Verari (probably more).
Dell Chairman and CEO Michael Dell: his company intends to double its lineup of AMD-based systems by year's end.
Jonathan Schwartz, Sun's president and CEO: his company is aiming to use the quad-core Opteron to double its AMD-based server business.
"The two companies said Lucasfilm would order servers based on the technology for the data centers at Lucasfilm that produce digital arts and entertainment. Lucasfilm said it uses AMD chips exclusively at its global operations at the Letterman Digital Arts Center in San Francisco and its campus in Singapore. The San Francisco campus alone houses a 13,500-square-foot data center complete with a render farm, file servers, and storage systems."
I don't for a second believe AMD has clock speed binning problems when they just managed to release a very old X2 3200MHz that overclocks to 3500MHz stable, but I DO have reason to believe they have low yields on the new chips fitting those same 2xK8 type of 65nm chip upgrade (K10) into the 120W/75W TDP they would like and have touted for long being the key. That's exactly why I believe the new boards will help, and why HT 3.0 with high speeds is not being utilized on these chips now (because higher HT/IMC frequencies can be power hogs).
I really am not and cannot draw conclusions for any desktop parts based on any chip released yet. And I do think Phenom FX is a different fish to Opteron, however slightly.
The marketing team of Intel is spot on with winning fans, praise and leadership: they launch at high as frequencies as possible, and make their minimum and maximum frequencies at launch quite high with immediate effort on clock rampings back and forth (along with price :D) ASAP. That's the way to qain a market share, to make a loud buzz from the initial launch performance, as that shock really does sink into and capture people for a while - generating knock on effect legacies. AMD needs to do something similar. And from what I'm seeing, they are working solid to launch higher clocks along with a good platform.
BTW, sorry about this again :D but that 30k 2k6 Inq, I feel pretty sure its 2x Phenom 3GHz chips there on that RD790 board :yepp: (yes I know that it states of single socket)
These Opty numbers should be trusted for clock scaling at its peak for future K10 in server benches IMHO. I wouldn't count on more than this at all (not sure if the Xeon numbers are true though)->
http://img525.imageshack.us/img525/3...naspecish7.jpg
http://img91.imageshack.us/img91/125...naspecfie8.jpg
It was 3 way crossfire, people that was there saw the system and ahd 3 cards stop the hype and miss information.Quote:
Originally Posted by KTE
One more thing those systems were bios modded for marketing reasons i know this for sure ;)
About K10 im not un informed. Please explain how Anand got a B2 chip that is somehow bad since you confrimed B2 was the good one.
OK metro.cl. I have your word against someone elses whose also a hardware website owner and who I trust just as much. This confuses me. So I'll forget about it until we get the systems to run tri-fire or whatever config and see the scores. Either way, it wasn't what they posted or claimed. :D
"About K10 im not un informed. Please explain how Anand got a B2 chip that is somehow bad since you confirmed B2 was the good one."
Me? I haven't yet said anything about Anandtech chips, because I didn't know. Dave has though and since it comes from AMD rather than speculation, I have no reason to distrust him. :)
The waters always clear away sooner or later, but benefit of the doubt has to be given to someone with a safe reputation, trustworthy and known honest as I feel, unless you have equal value proof otherwise (like another AMD rep saying something opposite). In which case, its better worked out and reconciled tamely in private IMHO. :)
Tell you what - why don't a few guys on here who doubt what Dave says, or one of their many online friends, ask AMD or their AMD contacts?
AFAIK, a 2p rd790 board isn't yet finished. If it was even close to completion, I'd be making calls like a madman trying to get that board on my desk ;)Quote:
Originally Posted by KTE
On another note, I've been staring at this Barcelona system for the last hour or so and decided that I'll just run the one "good" cpu for now....at least we'll all get some real numbers to clear up all this speculation and bickering...
Finally some action :).Quote:
Originally Posted by s7e9h3n
Tell us the stepping and the date of the BIOS :).
Btw, talking about the "2S systems"...
have anybody seem this one: http://www.theinquirer.net/?article=42307
:( :cry:
Anyone closer to the sources could please thell us that this is a lie? :(
Thats good news, Finally some numbers everyone here is interested in (unless you're gonna do a 500 connection sql test :) )Quote:
Originally Posted by s7e9h3n
and no it won't clear up the bickering, you know it will get worse :p:
first order of business
search for coldbug :D
This is good, I'm most interested in 1 X 4 core CPU on those x264 tests anyway. x264 automatically runs as many threads as cores.Quote:
Originally Posted by s7e9h3n
agreed...Quote:
Originally Posted by dinos22
Out of respect, could everyone stop fighting in the best thread on XS? Having said that, I dont think anyone went out of line here, but it doesnt do the thread good because some people could decide to stop posting. Now if that happens to be the wrong type of poster, we get no information anymore. And THAT will piss me off. ;)
PetNorth: interesting numbers. Did you do the research?
After reading this thread , i'm not interested anymore . I'm disgusted with this thread.
I'm with you, most of the people in this thread are worse than a bunch of ten-year-olds fighting on the playground. :shakes:Quote:
Originally Posted by [XC] Jaco
PM a mod and they will take care of this. Why stress ? :)
Guys, please respect s7e9h3n's thread! :up:
^ as the man says :D
Back on track!
qft++Quote:
Originally Posted by Ashraf
Here you go (sorry for the ugly pics : I broke my main camera and all I have to work with is this pos :p: )Quote:
Originally Posted by informal
http://img524.imageshack.us/img524/8162/bios1kh1.jpg
http://img209.imageshack.us/img209/2329/bios2yu3.jpg
And after messing around with the setup last night, it turns out that it's the board (or bios) which is causing the error in correct speed detection of CPU1. Although the cpus seem to detect properly before entering windows, there's nothing I can do to get CPU1 to read correctly once in. Hopefully, it can be remedied with a simple flash of the bios. Otherwise, I may have to get another motherboard which is easier than getting a new cpu :D
C'mon, Stephen, press any key and load windows :D
Oh yeah, forgot to add that there seems to be a problem with the memory detection as well. For some reason or another the 4x2gb sticks of PC5300 are running @ 159mhz SINGLE CHANNEL CL5. That's have a HUGE impact on performance. If anybody's curious, @ 800Mhz on the cpu and 159mhz on the mem, I can run a 39sec Spi 1m LOL :p:
So not even single CPU benches for us? :(
What is progress with L1N64 BIOS support for K10?
Anyway great job both s7e9h3n & Dave!
CPU1 won't run any faster than 800mhz. CPU2 runs @ the proper 2ghz, but on a dually board, you can't only run CPU2. If you're running a single proc, you can only use the CPU1 socket.Quote:
Originally Posted by Lightman
L1N64 Bios is going just as expected with Asus - absolutely nowhere....:mad:
So a K10 @ 800mhz and crappy memory gives 39s 1M ?
That doesn't sound very bad to me
hmm 39 sec spi 1M @ 800MHz CPU 159MHz mem, could full speed be a bunch faster than that, like maybe mid 20's? I know I say I don't care much about spi but I'm just curious.Quote:
Originally Posted by s7e9h3n
Quote:
Originally Posted by s7e9h3n
On the spi run, we're talking about 1 cpu running 2G and the other at 800 Mhz, with the memory at 159 single channel. Sorry just confused about what you ran that at.
Thanks for the pictures and latest info :up:Quote:
Originally Posted by s7e9h3n
Can you verify again the Spi 1M run on those 800Mhz and 159Mhz for ram :)?Since it is a LOT faster than kyosen's run on Akiba demo yesterday-49s@2Ghz :) (btw,thanks kyosen for those results :) ).I mean,is there a possibility that even if the reported speed is 800mhz,the CPU is actually running @2Ghz(and scoring the 39s)?Best way is to run Spi in dually config and bind it to CPU1(instead of CPU0) in process explorer.
Thanks in advance and good luck with the system :up:
Yeap, that would be my guess too. Coolaler has a 39s 1M run at 2G :)Quote:
Originally Posted by informal
From when did we start saying please 0_o.Quote:
Originally Posted by Ashraf
For all those gentlemen who have already decided not to buy AMD this winter, the Intel section is only a centimeter below the link for the AMD forum.
Perkam
Steven that's exactly the same time that 2GHz quad core Opteron gave on the Coolaler forum in SPI 1M (what we were all shown).
Hmmm.... there was a massive hue and cry about that on both sides. Seems that must have also been a buggy board/bios/chip. This is getting interesting. :D
On another note, gack, check the amount of problems you've had so far. Is this a first? :(
Having solid platforms for your CPU is something that any firm must ensure. I guess that's why Phenom is delayed too, and I suppose rightly so.
holy cow! The plot thickens.:D
{hits refresh button}
you suprised yet? :)Quote:
Originally Posted by s7e9h3n
we're trying. I can run superpi tonite. i do actually need a download list for me to run.Quote:
Originally Posted by Lightman
dear god...that BIOS (which looks like it's for an nVidia based board as my 3992 uses an Award bios) looks.....interesting....Quote:
Originally Posted by s7e9h3n
in any case, i can see what i get tonite.
thanks for the tips.Quote:
Originally Posted by informal
dave
Dave for me please:
- Cinebench 9.5 x64 Cinebench
- 3DMark06 CPU Test
- Everest4.0 Latency tests and FPU tests :) Everset Trial
- mBench memory subsystem benchmark mBench.exe
- anything else you will like (including server benchmarks, kernel compilation, etc...)
THANKS!
Dave, BA or still the same B1s?
a Cinebench 9.5 and 10 both on win64 would be great!
Oops....I was assuming one of the logical procs in socket #1 was running the bench, when in reality the process initiates with cpu1 and then is immediately handed off to cpu6 for some odd reason. Bound the bench to Cpu5 and it results in a 1m ranging from 39s - 41s. The memory is STILL being detected @ 159mhz single channel though. I tried to install Sandra to check other stats on the board, but it flat out wouldn't even run after installation. Upon reboot, I can't even get back into the OS. The rig still posts, but just doesn't enter the OS. When I try to re-install windows, I get a BSOD in the middle of the setup process. I'm tired and frustrated with screwing around with this config which AMD had claimed should work without a problem. I'm going to have this setup swapped out for another which I should have in a couple of days.......Quote:
Originally Posted by informal
On my 37" Aquos, it's more than 1" below the AMD link, making it too far for my mouse to wander :DQuote:
Originally Posted by perkam
LOL that was great!Quote:
Originally Posted by s7e9h3n
Such a shame s7e9h3n. Was really looking forward to some benchies. Looks like I'll have to wait till you get another rig next week. Lets hope it is a bit better :)Quote:
Originally Posted by s7e9h3n
Maybe 3DMark01? :DQuote:
Originally Posted by Lightman
Does 3dMark even install on x64 based OS's? I've tried to install PCMark on my 2x2218's running 2k3 x64 and it won't work. The only way I can run PCMark is to use a separate XP 32-bit install.Quote:
Originally Posted by massman
Have you tried running the installer in XP SP2 Compatibility Mode? Might have to do the same with the .exe once it's installed.Quote:
Originally Posted by s7e9h3n
maybe I'll give this a try myself, haven't run that in years
It's running very well under XP x64 in contrary to stupid PCMark...Quote:
Originally Posted by s7e9h3n
Ahhh, ok....good to know as I should be getting an 8800ultra (or two) early next week.Quote:
Originally Posted by Lightman
OT: Noticed you have a T-mobile MDA in your sig. I've got the MDA pro and when I get bored of messing with computers, I just OC my phone :p.
s7e9h3n & dave_graham
SiSoftware Sandra: http://www.sisoftware.co.uk/index.ht...buy&langx=en&a
-Arithmatic, Multimedia, Multi-core efficiency, Mem Bandwidth, Mem Latency, Cache and Memory Latency benchmarks.
Include EVEREST Memory Read/Write bandwidth please.
CrystalMark: http://www.softpedia.com/get/Tweak/S...stalMark.shtml
- ALU, FPU, MEM test. Not bothered about HDD, GDI, D2D, OGL.
wPrime: http://www.wprime.net/download.html
- 32M and 1024M if you overclock.
WinRAR: http://www.rarlab.com/download.htm
- Tools> Benchmark and hardware test. :)
ScienceMark: http://www.start64.com/index.php?opt...d=72&Itemid=76
- (couldn't find any more recent x64 Windows edition) File > Run all benchmarks. :)
TrueCrypt: http://www.truecrypt.org/downloads.php
- Tools . Benchmark > 100MB and 1GB buffer size.
You can do PassMark two CPU and one Mem test aswell: http://www.start64.com/index.php?opt...id=34&Itemid=1
SuperPi 1M and 32M would be good too for comparison please. :)
Just run whatever the damn you can :D I'd ask for server benches but I damn well know how costly they are. All above should run on x64 Windoze though and gives us a good allround indication.
Gaming benchmarks? That wouldn't interest me unless you wade through which games at what stage/settings are CPU bottlenecked and then test to compare. Cheers.
EDIT:
I forgot POV-Ray rendering: http://www.povray.org/download/
QX6850 gets 98 seconds in 3.7 and E6850 gets 196 sec.
Valve Map Compilation, Particle Simulation bench would be good.
If you can, run a single/dual core only to see the power and scaling of that in comparison to quads/octal. News outlet bloggers are going on about Core 2 being more powerful as in single core efficiency because of its 4-execution engine, so I'd like to see how this compares.
As always, thanks!
OT: Heh! I was OCing my previous MDA based on OMAP CPU (from 180 up to 240MHz :D ) mainly because of Worms WP but with this new one I can't find app which will run slow on it :p: !Quote:
Originally Posted by s7e9h3n
BTW what is powering your phone and how you OC it?
Back on topic: So you'll get Ultras together with new (hopefully working) board for K10! Looks like it will be interesting week :) . This weekend I will chill out watching F1 on TV so no rush with testing please :p: !
It was a joke of course!!! Bench new stuff 24/7 or die!! ;)
Just to clarify it for others:Quote:
Originally Posted by PetNorth
What he meant by performance scaling, is processor frequency scaling efficiency.
i.e. performance change/speed change x 100
Speed bump was 25% so maximum expected linear scaling is 100% of that 25%.
OT: Is it one of the HTCs?Quote:
Originally Posted by Lightman
yes ;)Quote:
Originally Posted by KTE
I decided to give this system one more shot at booting up before having to trade her in. To my surprise, I was able to get into windows on the first attempt. I installed a newer version of Sandra and ran a couple of the benchmarks. Needless to say, the results seemed WAAY off base as Sandra had this Barc. setup ~400% slower than a dual Xeon 3220 system. I've got some SS of some other slightly less embarassing benches (but not by much) which I'll post later as I'm writing from my phone ATM.
OT: @ KTE and Lightman - my phone's a qtek9000 -> a rebadged HTC Universal powered by an Intel XScale 520Mhz cpu ;)
That's quite sad. Now I just want to see when the first OEM K10 system is up for grabs and if that suffers from the same problems. Its kinda obvious K10 is not in platform stability mode yet. :( Whether its the chip, MB, BIOS, can't really tell from here.
Worst thing is, everyone can get differing results and not know if the results are accurate, underperforming 1%, or by 500%.
s7e9h3n: Is there any way to underclock the CPUs through BIOS? If you do that, like to 1.5GHz, would they run?
I'm thinking maybe the earlier chips have clock instabilities here. That's usually how a failed overclock reacts.
Ahh well, let me make note of a few related things I've seen around, just to show you what can happen. Just be careful that these server benchmarks vary very widely depending on how optimized you tune the software running (and in this case, your BIOS/CPU rev)
Techreport benching: http://www.techreport.com/articles.x/13176/4
Anandtech benching: http://www.anandtech.com/IT/showdoc.aspx?i=3091&p=5Quote:
Originally Posted by Opteron 2347-2350
Specjbb2005
TR: Dual Opteron 2350 (2GHz) = 88949 Bops
AT: Dual Opteron 2350 (2GHz) = 70363 Bops
TR: Dual Xeon E5345 (2.33GHz) = 87718 Bops
AT: Dual Xeon E5345 (2.33GHz) = 68303 Bops
Get me? :confused: And the SPEC submitted results are usually higher than these, especially for Xeons they are submitted as the authors mention, with all tweakings and so on.
under windows xp pro, sp2, on a 3992 (single plane) with 2 x 2347 B1 processors and 4 x 1GB DDR2-667 (CAS 5) ECC/Reg modules:Quote:
Originally Posted by KTE
8 total threads
32m: 11.406 seconds
1024m: 364.062
4 total threads
32m: 23.015 seconds
1024m: 742.875
1 total thread
32m: 84.578 seconds
1024m:
also, CPU-Z is inaccurately reporting clock speed but wPrime IS accurate. weird, i guess.
I'll have the untouched WPS result files available after everything is tested.
see below
i have 2 x 2344 (HE processors) rated at 1.7ghz that I can test against as well.Quote:
Originally Posted by KTE
dave
You mean as if they were throttling???Quote:
Originally Posted by KTE
I don't think AMD CPUs can do that.
AMD processors make use of dynamic multiplier manipulation to enter into different power stages...it's MUCH more flexible than Intel's currently (which, iirc, is 3 different states).Quote:
Originally Posted by Piotrsama
cheers,
dave
Here's a few Sandra shots. I guess I didn't save as many as I had first thought :rolleyes:
Here's what Sandra says about the entire config. Note how off some of the numbers are:
http://img513.imageshack.us/img513/4...drainfotq0.jpg
Here's the CPU Info:
http://img124.imageshack.us/img124/8...acpuinfud4.jpg
Memory Bandwidth: Note the Bandwidth efficiency :shrug:
http://img513.imageshack.us/img513/3...drabandgw7.jpg
Here's the Cache and Memory bench:
http://img509.imageshack.us/img509/2953/sandracmdn8.jpg
One thing weird I did notice with these cpu's is how they're seemingly handing off workload to each other. If I don't set the affinity for a bench, say like Spi, the cpu's will randomly hand off the calculations one another every so often. It's quite odd and it also causes the bench to run slower. In all of the Sandra benchmarks I ran, it seemed that only the cache and memory bench from above gave the cpu's a consistent workload. With the other benches, the cpu usage was erratic and unpredictable. I guess it's a possibility I have a bad install of windows, but hopefully with some new HW, this will all be sorted out.....
Dave: Have you installed just windows on that? If linux, would mind in making some trials for gromacs gcc compiled benchs, with the assmble loops on it (specfp disable those).
If possible, let me know, and send you a few "guidelines". ;)
Yep, Yep!!
Looks like Barcelona's bitten the BIG 1....
So much for AMD!
We need desktop parts, and pronto.
Perkam
It's interesting how things turn sour as soon as a couple of members that regularly post in AMD section joined here :rolleyes:
showoff :p: :D j/kQuote:
Originally Posted by s7e9h3n
what size screen is the phone and how many hours did it take you to write a couple of sentences :D.....i tried surfing the net on a phone once and it was the most frustrating experience i've ever endured lol
Perkam yeah desktop parts PLS :D
Quote:
Originally Posted by [XC] M411b
WAIT WAIT.....are you being sarcastic XC? lolQuote:
Originally Posted by perkam
I hope so cuz aint that cache and memory bench like AMAZING?...or am I reading the numbers all wrong...lol
and perkam you saying that in desperation or in seriousness cuz of the excitement of the above performance...rofl
Im confused!!!!
I think that bottom screenie is AMAZING............
1573 kb/sQuote:
Originally Posted by KTE
0 errors
over 1.1GB processed
12:02 minutes
interesting thing to note about this is that a Via EDEN could probably do a GREAT job on this one.Quote:
Originally Posted by KTE
Buffer size:
200mb:
Encrpyt, Decrypt, Mean (mb/s)
Twofish: 46.7, 48.1, 47.4
AES: 46.3, 40.5, 43.4
Serpent: 35.2, 35.7, 35.4
AES-twofish: 23, 21.6, 22.3
Twofish-Serpent: 20.5, 20, 20.2
Serpent-AES: 20.3, 18.3, 19.3
Serpent-AES-Twofish:14.7, 13.5, 14.1
AES-Twofish-Serpent: 14.4, 13.5, 14
1GB:
Encrpyt, Decrypt, Mean (mb/s)
Twofish: 46.8, 48.2, 47.5
AES: 46.3, 40.6, 43.4
Serpent: 35.3, 35.7, 35.5
AES-twofish: 22.9, 20.8, 21.9
Twofish-Serpent: 20.2, 19.4, 19.8
Serpent-AES: 20.1, 18, 19
Serpent-AES-Twofish:14.5, 13.2, 13.8
AES-Twofish-Serpent: 14.2, 13.1, 13.6
Yes it is good but the screenshot above it seems to contradict it if I'm reading it correctly :shrug:
And you're basing this all on a sandra screenshot of all things? That's bright. :rolleyes:Quote:
Originally Posted by [XC] M411b
with a terribly buggy bios at that...Quote:
Originally Posted by rogueagent6
Thanks for the effort guys! :) Must be crazy getting these working...
BTW fellas (everyone else), the results weren't amazing. The results were poor actually :( and obviously at least some were very buggy too. Here's what I've heard:
4 steppings so far - B0>B1>BA>B2 BA has the B1 fix. I'm hearing there should be only two steppings out to reviewers, the last two. Can you guys please run CPU-Z?
Dave I know you said you have a B1 but I'd like to see what CPU-Z reports according to core register values please. wPrime reports false frequencies many times, and usually I've seen it report BIOS frequencies and not current CPU frequency. But CPU-Z will report the current frequency given by PLLs/core. If you have buggy/faulty or unsyncd MB PLLs or BIOS, the frequencies will change and be messy (but CPU-Z will report that real-time).
What frequency do you guys get if you run this utility? http://web.inter.nl.net/hcc/J.Steunebrink/chkcpu.htm
I think I know how to recognize the stepping in CPU-Z:
http://img261.imageshack.us/img261/4893/b21cp1.png
http://img480.imageshack.us/img480/3750/b2yh2sw5.png
F-2-2 : B2 core
F-2-A : BA core
That's as far as I can work out from AMD K10 PDFs. I know for sure that above screenshot is a B2 core BTW.
s7e9h3n/dave_graham: There's also something seriously wrong with the results/info there, your mem bandwidth and conflicting info in Sandra for one and then comparing Daves results around with those before, like those on Coolaler forums- they all differ. For reference, I ran this about 5 months back on a single core 3.4GHz P4 DDR400.
http://img208.imageshack.us/img208/2996/3ghzp4ug2.png
An Conroe 2.4GHz nicely beats that score. :) How can these quads get 5.9 in WEI if the encryption/decryption is so slow while very similar to one of the tests ran to determine the end score, and that P4 doesnt even get 3.8 in WEI?
Dave, the 1573KB/s for WinRAR... well E6320 at 2330MHz gets 1302KB/s.
Also, Steven, your memory bandwidth score was the same as a Pentium 4 820 D on Intel 955X chipset with CL5 2x1GB PC2-5300 RAM. :confused:
AND your cache/memory speed was the same as a P4 3.4GHz while your bandwidth was less than a 2.4GHz Q6600. For reference, the Q6600 2.4GHz on Intel BadAxe2 2x1GB PC2-5300 gets 41502MB/s.
And Dave, that wPrime....
Coolaler forums and previous told scores (IIRC it was Moviemans Xeons):Quote:
8 total threads: 32m: 11.406 seconds - 1024m: 364.062
4 total threads: 32m: 23.015 seconds - 1024m: 742.875
1 total thread: 32m: 84.578 seconds
2GHz Opteron 2332: 32M Test 10.641s (8 cores)
2GHz Opteron 2332: 1024M Test 327.437s (8 cores)
2.66GHz Xeon X5355: 32M Test 8.203s (8 cores)
2.66GHz Xeon X5355: 1024M Test 243.812s (8 cores)
WOW this is all over the shore :shrug: :shakes: .... What Gary of AT said 30-08 comes to mind:
Keep it up! Lets hope some decent BIOS and fixes are made fast.Quote:
"The latest Barcelona chips are B02 steppings with one more to go. Believe me, the reason we did not post any numbers at Computex or since then is the simple fact that the CPU/boards/BIOS have undergone dramatic changes over the course of the summer. If you have an earlier stepping there is a very good chance that HT and the secondary cache is disabled, this will affect the benchmarks dramatically. We expect to see final stepping chips and board revisions early next week, until then, it is all speculation for the most part.
The one caveat that I will add, this chip really does not get into a groove until you get over 2.4GHz and then it scales incredibly well. Also, the first RD790 boards we have will undergo another spin so any Phenom results with those boards are subject to interpretation depending on whether you like AMD or not."
So the cpu is set at 800MHz with the memory @159 and single channel cas5? I understand this well? This means that is really fast, even in Spi. If it would scale just as the K8 this should get it to a max of 30 sec with cpu @2GHz and memory @667/128bits. The platform is really buggy.Quote:
Originally Posted by s7e9h3n
Thanks for sharing!:up:
Later on, he posted this:Quote:
Originally Posted by planoru
SuperPI was running on a 2 ghz core.Quote:
Originally Posted by s7e9h3n
LOL, your the only one that understood that I guess! :up:Quote:
Originally Posted by GenTarkin
AMAZING isn't the word for those numbers...:clap:
You're right I am basing that off of one Sandra screen shot. Those numbers are huge, and I was being Sarcastic. Didn't realize how serious peeps in this thread are getting, sry.Quote:
Originally Posted by rogueagent6
http://img177.imageshack.us/img177/884/cpu1my4.jpg http://img297.imageshack.us/img297/3861/cpu2dj3.jpgQuote:
Originally Posted by KTE
This is what happens when I try to validate it:
http://img502.imageshack.us/img502/2551/nonvalidmt4.jpg
False or True Frequency? :p: Answer below ;)Quote:
wPrime reports false frequencies many times, and usually I've seen it report BIOS frequencies and not current CPU frequency. But CPU-Z will report the current frequency given by PLLs/core. If you have buggy/faulty or unsyncd MB PLLs or BIOS, the frequencies will change and be messy (but CPU-Z will report that real-time).
http://img502.imageshack.us/img502/8236/wprimeck1.jpg
I finally figured out what's REALLY going on with this rig. Thanks to AMD's monitoring software, it became quite apparent what's happening with the cores. Somehow, either the PLL or the bios has screwed around with the P-state of Core0, CPU1. A setting in bios allows the cpu's to enter the OS either in maximum performance levels or in power saving mode. I have it set for max performance. It works as it should, except for ONE minor little bug :rolleyes: which has resulted in one of the more interesting screenshots I've ever seen:
http://img134.imageshack.us/img134/8536/cpumo8.jpg
Core0 is STUCK @ 800mhz in max power-saving mode (as indicated by the bars on its left) while all the 7 other cores run @ 2G. The problem isn't with the cpu's since it happens only with Core0, CPU1 even when I swap sockets. I'm still wondering how it's possible to have only one core running continuously @ 800mhz and 3 @ 2ghz without crashing the OS. (and YES that core is not being mis-read as a Spi1m with affinity set to Core0 turns out a time of >1minute 30 seconds :zombie: ) This one's definitely going to have to be swapped out for a replacement......
You did just fine, I think most of us realize you were being sarcastic.
Yep, there was something definitely wrong with the bandwidth score. Check out the bandwidth efficiency -> 2%. Octal Opties are capable of some outrageous bandwidth numbers if everything is running properly. Here's what my quad 2218's (2 x 2218 DC) did on the Asus L1N64 (they're capable of more but I haven't benched them hard)....Quote:
Originally Posted by KTE
http://img179.imageshack.us/img179/2549/511memser8.jpg
have you tried crystalcpuid? can you change multiplier with this tools?
http://dl.crystaldew.info/download/C...PUID413x64.zip
look weird core 0 running at 800mhz while 3 other core running at 2Gig
Nope, haven't tried that, but I can't imagine it would work as core0 seems to be totally locked into that P-state. I think either the bios may be corrupted or the PLL is borked on this board....Quote:
Originally Posted by tictac
Yes, it really does look weird but it runs even weirder :p:
Thats some strange happening you have going there ??????
But good work showing us all this, Cheers
Se here on XS we could see new AMD feature called independent core clocking in action thanks to small BIOS bug :) .Quote:
Originally Posted by s7e9h3n
Look at this from the other side s7e9h3n, your saving energy! :cool:
FYI, i can not validate a s3992 board with 2210's with CPU-Z, seems to be more related to the board's chipset.Quote:
Originally Posted by s7e9h3n
These are amazing results.Quote:
Originally Posted by s7e9h3n
With Numa enabled i get 10559 MB/s(int) and 10560 MB/s (float) on the s3992 with two 2210's(1,8 GHZ) and 4GB Reg/ECC Ram running at ~300MHZ.
Without Numa 6013 MB/s (int) and 6004 MB/s (float).
Not sure but will this be any help with the oddly behaving CPUs
http://www.amd.com/us-en/Processors/..._15258,00.html
That tools got a CPU MSR Editor..Quote:
Originally Posted by s7e9h3n
in Function tab..
but first beside the cpu name change it to the Processor node 0, core 0
if it is P-State problem...
then open MSR Editor
Read register : C0010062
Bit : 2 - 1 (control the P-State)
set it to 0 (hexadecimal)
0h = P-State0 (Maximum performance)
i hope it helps...
Things get more interesting by the day. I think it is safe to say that the bios is so buggy and the boards suck so much still that it too early to make any concrete conclusions about Barcelona.
As for the Core 0 issue, this might be pretty cool actually because if that is true, perhaps it is possible to clock each core independently? That would be pretty fricken sweet for when you are pushing balls to the wall and one core is holding you back. Though I'm not sure how that would effect stability or anything....
Thanks for all your hard work and thanks for sharing your results so far.
Barcelona allows for each core to clock independent of each other. this was present in early design docs and is still true to this day. to take advantage of this, however, you definitely need the BIOS to support it and the processor driver within windows. Again, this is much more interesting test metric that, should S7 and I be able to figure out a good benchmark for it, would definitely provide how well the Barc. is able to do in power savings over Clovertown.Quote:
Originally Posted by stealth17
this has been confirmed by AMD.
DDR3 is not present on the cpu and will require a repackage of the core if/when it arrives (Dozer).
hope that clarifies a few things.
dave
Very good stuff guys. Keep up the good work.
I really hope AMD irons out problems (if they're with the CPU) before they transition to DDR3 and 45nm.
Quote:
Originally Posted by dave_graham
Yes it does, thanks Dave
Phil
Stephen, try changing p-states with RM Clock:
http://cpu.rightmark.org/download.shtml
Dave, could you run SuperPI to end the PI question?
http://www.xtremesystems.com/pi/
my running superpi won't end the PI questinos since i'm running B1's on a single plane board. that's not a fair test to ANYONE. :)Quote:
Originally Posted by doompc
cheers,
dave
this really isn't an AMD problem. it's a very poorly handled platform issue.Quote:
Originally Posted by freeloader
cheers,
dave