Printable View
s7e9h3n,
Thanks for the info concerning these new processors :toast: :toast:
Guys(Gals) like you make xtremesystems the place to be.
Concerning P95: What constitutes stability? -- Being able to use your computer the way you want to. I am a scientist, I use my computer for many purposes. I even use it to calculate Monte Carlo simulations concerning particle transport. It has not failed me yet even though it is not P95 stable.
Since the Venice cores are gona be awailibly for end of april in sweden. When is it gone be reachable in the states?
/Cherio!
Let the man get a rest, the next intresting test would be with a DFI kard. Hope he have one home. :
4th of april in us, I think Newegg will have them as one of the first in the us.
This post was back a few pages here, but I missed out on some posting during the St. Patty's evening.Quote:
Originally Posted by s7e9h3n
SO, allow me to retort. I have to disagree that it is going overboard because you can do a different type of testing. Prime95 tests are usually less BSOD prone because you aren't pushing the limit. You probably wouldn't even have to check it all that often... I log into my machine at home from work all the time ( I try to keep my bandwidth maxed out ). It is a little laggy trying to log into a prime95'in machine, but it can be done.
since i dont wanna filter through 17pages of benchmark requests and etc... what was the verdict on the venice chip with subzero temps?
Even when S&M makes it all the way through, OCCT dies on me w/in 2 hours :wierd: :bsod:
With temps going so low, Is it going to be worth getting an FX57 if you are not going to get the below 0c?
Most people with good air or water will be reaching the same speeds with a diego than with an FX, right? What would be the benefits on the FX on these new cores?
Note: Even though this a slight tangent, I beleive it pertains to the importance of the testing of this chip.Quote:
Originally Posted by tritium
I totally agree and think that most people know you can be clocked higher than prime stable and run other apps/games without a problem.
However, since we all know this we can add that factor to the baseline of a prime95 test. For instance: my winnie is prime95 stable at 2.4ghz 266 1:1, but i know I can run it at 2.55 for most apps without crashing. It is by far easier to use prime95.
The more varied types of tests used the better, IMO.
:D AFAIK the FX-57 will have also the 0.09 SAN DIEGO core, but more cache so if You have a big wallet go for it and get it on single phase.........Quote:
Originally Posted by REDKEN
:toast:
It's called xtremesystems.org cause that's shorter. Period. :toast:Quote:
Originally Posted by HaLDoL
wow great job s7e9h3n keep it up! :D
Sorry for double post mods, but we need to straighten things out here. Venice is E3 with 512KB L2 cache. (1=128,2=256,3=512,4=1024,of which the lower are for semprons and the E is for revision E w/ SSE3, tampered mem controller (we hope) and w/out the sub-zero bug.) The E4 "San Diego" has 1MB L2 of cache - no matter whether you go 4200+ or FX-57. It *might* also be that diegos are gonna use sSOI (Strained Silicon On Isulator) production process, which allows for higher clocks (especially sub-zero, or so I've heard).Quote:
Originally Posted by HARDCORECLOCKER
The big thing about the FX-57 though (compared to its diego sisters) is that it's the cream of the puff - they're binned the highest of all, also they've an unlocked multi (which is could turn out useful in certain cases). Also, AMD are usually extra generous when it comes to the warranty of the FXs - if you pay the price you can also expect the most. Other than that, any locked diego would probably be fine as well.
(Not all of the above stuff is confirmed, feel free to flame.)
:D SRY - if You are right with Your "E" theory it was my bad - then the cache might be the same.......... :toast:Quote:
Originally Posted by krille
That seems only logical in this case. However, I'm quite sure about the numbers just can't remember from where I got it. Let's hope for the sSOI though! :DQuote:
Originally Posted by HARDCORECLOCKER
i tought that sSOI was just first for .90nm FX and then they will use it for other chips, seems logical because fx are top of the line and the fx 57 first 90nm production fx will need the sSOI to scale further.
but im not 100% sure where i read that and they said that soi was limited to 3.0ghz and that is why they go for sSOI
No, "sSOI" was first incorporated in a few top 130nm parts (FX-55 and 4000+), as well as all 90nm parts.Quote:
Originally Posted by leviathan18
DSL (2nd gen strain) is going to be worked in to certain (maybe all) 90nm lines throughout H105.
I don't think they ever said that SOI was limited to 3GHz, but strain certainly does help with power and speed.
Right now all I have to say is...I miss a new post from s7e9h3n ;)
ok, so lets host this sight on a server that isnt stable and crashes every 20 min, and then we'll see who cares about stability :stick:Quote:
Originally Posted by HaLDoL
GOOD LUCK s7e9h3n, your a swimmer with a paper-cut in a tank of STARVING sharks :hehe:
Nice info on the new CPU but I would like to know also what your best prime stable is...just for an hour not some 24 hour marathon....usually anything flaky dies in under 10 minutes and yes I can run some things higher than prime stable but I like knowing if something messes up that odds are it's not my OC....I hate BSODs and am not into suicide shots myself but to each his own. I'm more impressed by 24/7 numbers.
first post on this forum, 4th or 5th overclocking forum i'm registered with
first off, s7e9h3n, you and i need to be friends.....
great thread u got going here, keep it up, can't wait to see your 4x256 results, 3500+ venice results, and memory controller results
alot of you are speculating on the PR releases of the venice/san diego, i personally check http://www.c627627.com/AMD/Athlon64/ to see what's coming out/when/with what, don't know how accurate it is, but it sure is purty
hope to have many good conversations/debates w/ all of you :toast:
Welcome to xtreme, Molester! :up:
I am also eager to see results with 4 sticks and 1T :)
4x256MB BH-5Quote:
Originally Posted by SAE
:slobber:
Welcome Molester, nice name there.:D
And we're all dying to see 4xDIMMs!
Cheers! :toast:
Wow, I don't check this thread for half-a-day and I now can't find where I had left off last time. I didn't really feel up to a hard-core benching day yesterday so I basically ran a few tests here and there. First thing I did was try 4x256Mb Bh-5 on the board. My current board DOES have the cold-boot issue and a non-start initially isn't a big deal. But I ended up sitting there trying for about 5-10minutes to get to the bootscreen - NO GO. I don't know for sure, but it seems as if there's a problem (maybe with my neo2) with BH-5 and this setup regardless the number of mem slots occupied. I was able to boot with 2x256, but this even took longer than usual. Trust me, the bh-5 is able to perform once booted, but it acts oddly with the setup. Here's the biggest issue I can't figure out - Check out this screenshot and tell me if you've ever seen this happen: :rolleyes:
http://img115.exs.cx/img115/6046/cpuzbh55bw.jpg
i think if you pay I.E 180$ for a 3200 and yo can take it to 3.2ghz stable then you will see stabilty matters because you are getting the most of that 180$ that you expend in the cpu and you have a nice upgrade made by OC not everything is bigger numbers :P
(is so difficult 4 me when i have to write in past future and present in the same sentence just dont mind my mistakes sorry :S)
Oh, and I haven't yet got a chance to read over all these new threads, but from the looks of it, you guys are talking about stability. Like I said in one of my first posts, I'm not here to "prove" the stability of this cpu @ a certain speed. I just wanted to show it's capabilities. Last night, however, I did run 3dmark01 @ ~3100Mhz and it passed, no problem. I didn't save it, however, since the the driver didn't load correctly in windows and for the fact that I had a crap video card. Now that I think about it though, I guess scoring ~19,000 with no tweaks on an old 9800 pro isn't all that bad.
well it reports the correct amount of ram... maybe cpu-z just needs to be updated?Quote:
Originally Posted by s7e9h3n
Well, I guess you can say that. I'm running 2x256 in slots one and two. But as you can see in the schot on the right, it detects 1G of mem... :confused:Quote:
Originally Posted by ozzimark
:slobber: That's interesting - AMD always denied using strained silicon yet, but reading Your post I remembered one of my screenies - and dang ! - look here:Quote:
Originally Posted by terrace215
http://img8.exs.cx/img8/38/ssoi1wi.jpg
Think I will start another thread to see if all FX-55 will be recognised as sSOI by CBI....... :toast:
I think the issue of BH5 is due to the board not like BH5. Thats why for a while, everyone was sellin their BH5's. After the NF4's came out, the supply dried up.
Here hardcore ;) :Quote:
Originally Posted by HARDCORECLOCKER
http://img17.exs.cx/img17/457/newspi1xa.jpg
@ 3100mhz 3d mark stable that is nice :D
thx for all the info you gave to us :D
Well, at least there's no problem with bandwidth:Quote:
Originally Posted by situman
http://img115.exs.cx/img115/9925/bh52657au.jpg
http://img115.exs.cx/img115/9205/2754wx.jpg
:D
I just used the program and mine says SSOI 0.13.Quote:
Originally Posted by HARDCORECLOCKER
funny though i cant get this chip stable past 3.2 gighertz
oh well, my theory is AMD is putting all the good stuff in the new chips coming out next month.
I just bought my fx55 last month mines like week 50
:D THX so far........ :toast:Quote:
Originally Posted by s7e9h3n
So then run to the second test of Prime95... ;) That's showing it's true capabilities. :cool:Quote:
Originally Posted by s7e9h3n
:YIPPIE:4x256 (DS, right Stephen?), 275, 2-2-2-5, 1t!!! Looks like the ODMC really rocks :bounces:.
Hey wicked, I tried to run P95 blend @ 3.2Ghz, but it could only get to about 1.5 hours then returned an error. Still, I didn't even expect it to get that far.Quote:
Originally Posted by WiCKeD
I think everybody is misunderstanding me. I WAS NOT able to boot with 4x256. Only 2x256 worked. It may have just been due to the fact that I had just switched out my 2x256 TCCD and not enough voltage was supplied to boot at ~225FSB. The CPU-Z is somehow showing the correct memory slots populated, but is erroring in total amount. Maybe I'll try to boot with the 4x256 now that I've updated the bios to the more friendly 1.37mod. Those screenshots are of 2x256 since I figured I had my bh-5 out already, I'd test them out with this cpu.Quote:
Originally Posted by Vapor
Even 1.5 hours at that speed is impressive...that's good enough for most gaming then IMHO. That's awesome :toast: . Thanks for trying Prime at all in this forum of suicide hungry OCers
I'm really dying to see 4X 512MB double sided module run at DDR400. If those 256MB modules are double sided, please try and get them to run to see if AMD actually made this improvement in the memory controller.
Hmmmmm....darn, thought that was with 4xDS. Looks like you might have even more reason to get the DFI board now.Quote:
Originally Posted by s7e9h3n
Anyway, still quite impressive....keep up the good work.
Damn! If only I had a fifth slot for my booster :D ....................
Caution. The DFI currently has no official/beta bios out to do dual channel using all four slots!
There's an alpha already working great (I already tested it) - it's 3-16 - but for now you have to stick with slots 2&4 :D
SAE, what you're saying is that there isn't even a good board out now that can even use all four slots with BH-5 (with the DFI's support MIA and the MSI going bonkers on startup)???
for the timings these people are asking of this stuff, I doubt we'll see 4x bh-5 anytime soon.....Quote:
Originally Posted by Vapor
anywho, all fx-55's are sSOI.... Just thought ya'd like to know. :p:
Understandable :) Thanks for all you have done, it is appreciated :toast:Quote:
Originally Posted by s7e9h3n
Could you please clarify at your leisure? :confused:Quote:
My current board DOES have the cold-boot issue and a non-start initially isn't a big deal. But I ended up sitting there trying for about 5-10minutes to get to the bootscreen - NO GO. I don't know for sure, but it seems as if there's a problem (maybe with my neo2) with BH-5 and this setup regardless the number of mem slots occupied.
Is it a cold-boot issue only with the BH-5, or the old winnie cold boot issue no matter what memory is being used?
Ok, these are probably gonna be the last screenshots of any benches for this chip. I'm gonna take her off my mach 2 since she's spent more time on it that my FX55 :rolleyes: I was able to pull this off last night:
http://img214.exs.cx/img214/9666/spi33062dn.jpg
I'm positive that the 3800+ is capable of much more than I showed, but I really don't have the time to tweak the settings. It seems as if the cpu is slowly starting to get burned in as it's ability to OC has been increasing gradually. For all you suicidescreenie junkies, here's the mother of all screenshots (thanks to ozzimark for showing me f5 technuique for cpu-z):
http://img217.exs.cx/img217/304/34006am.jpg
Just wanted to apologize for the perma ban threat last night. Nobody gets a perma ban on a whim, they only happen for very specific reasons and I was just spouting off. I'm taking a break from the forum for a few days, I've been wound too tight with everything going on in my life :slapass:
Please don't harass s7e9h3n and other informants. They bring us behind the scenes info that we would not have access to otherwise.
The FX-55 is strained silicon-on-insulator .13um, it was AMD's first mass-produced strained silicon chip.
No, it's a cold-boot issue with every and anything. My board simply just doesn't boot on the first attempt at most settings. I have no idea why since my last neo2 had no problems (well it didn't, until it blew itself up :D )Quote:
Originally Posted by EMC2
dayum...19 pages of replies in 3 days
not bad!Quote:
Originally Posted by s7e9h3n
whats the f5 technique?
Hate to, but... trying to make sure :) Hope you dont mind...Quote:
Originally Posted by s7e9h3n
So as far as you know, it is an issue with your MB, not the Venice? (as in your FX has the same problem?)
Again, thank you very much :D
No prob, definately the mobo. Now that I think about it, I'm probably gonna go buy the dfi tonight :DQuote:
Originally Posted by EMC2
i'm assuming you've tried various bioses? or do you have to use a specific one to get the venice to work?Quote:
Originally Posted by s7e9h3n
I've tried them all. Of note: the Venice appears to function just like any other s939 cpu with this mobo (granted I only tried bios 1.36mod and 1.41mod with her installed). I don't think that there's going to be an issue with incompatibility with the new cores.....Quote:
Originally Posted by ozzimark
Good news! :toast:Quote:
Originally Posted by s7e9h3n
Hey s7e9h3n, thank you so much for all the infos you brought, it's been a very interesting thread so far.
Have you decided yet if the 3500+ would get the same treatment?
thanks again
Yeah, thanks for your time in this thread s7e9h3n!
I ask the same as LowRun, do ya have time to try out the 3500+?
I leaning towards no, but I'll think about it.... ;)
I would be happy with just runs on air, it doesn´t have to be such a HUGE thread as this :)Quote:
Originally Posted by s7e9h3n
good to hear! hopefully the turion will have the same fate for the neo.
and because i missed before..
open cpu-z. hit f5.Quote:
Originally Posted by stealth17
note that there is now a bitmap in the folder cpu-z is in. it's much easier to get a screenshot this way. :toast:
Stick the 3500+ in IMO :)
As you feel man, it's allready been some very good infos there and i'm happy with that :DQuote:
Originally Posted by s7e9h3n
I have to thank Cathar for making me wait for his G5 waterblock, otherwise I would have bought a crappy winnie setup. Now I have the option of waiting for Abit to get its act together, DFI to solve some hardware issues and ATI's Bullhead to come out and ofcourse the Venice. Gonna be very interesting in about a month.
Hello.
s7e9h3n and I have agreed to consolidate the most frequently asked questions regarding the 3800+ Venice and compile a simple FAQ. The contents of this post has also been appended to the first post in this thread. Any and all updates will be added to the first post should the need arise.
Kindly consult with this FAQ on a regular basis so as not to ask questions that have already been answered. I have also added links to all of s7e9h3n's screenshots for those of you who find it tedious to search through the hundreds of posts in this thread. This should come in handy for the lazy folks, myself included. :D
Cheers!
:toast:
POST #1 to POST #194
Q: What are the major components being used in the system you are testing this CPU in?
A: A64 3800+ Venice + AMD FX-55 HSF
MSI K8N Neo2 Platinum (Bios 1.41mod)
G.Skill 4400LE (2x256MB)
Antec EPS12v True 550 (modded 3.3V, 5.0V, and 12V rails)
Q: When can we expect to see this CPU in retail channels?
A: You can expect to buy these through retail channels starting first week of April.
POST #195 to LAST POST
Q: What are the major components being used in the system you are testing this CPU in?
A: A64 3800+ Venice + Mach 2
MSI K8N Neo2 Platinum (Bios 1.41mod)
G.Skill 4400LE (2x256MB)
Antec EPS12v True 550 (modded 3.3V, 5.0V, and 12V rails)
Q: What are your loaded and unloaded temperatures?
A: -29C Idle and -15C Load. Measurements were taken from a thermocouple attached to the side of the IHS of the 3800+.
Q: How does the 3800+ Venice compare with an FX-55?
A: FX-55 (11x291MHz) vs. 3800+ (11x290MHz)
Q: Do you have any results that would give us an idea of how SSE3 affects performance in applications that are capable of taking advantage of SSE3?
A: SSE3 performance screenshots.
Q: Have you noticed any peculiarities with regard to the behavior of the 3800+ Venice?
A: This cpu, for some reason, doesn't like its default multiplier (12x) at higher frequencies. It requires much more voltage at 12x to run any kind of benchmark. On the other hand, 11x seems to be much more of a "sweet spot" @ 3.0Ghz and beyond for this chip. It seems to run smoother and with less heat and vcore than both 12x and 10x.
Q: What is the default VCORE for the 3800+ Venice?
A: 1.4V.
Q: What is the highest frequency you have run 3DMark2001 SE at?
A: Approximately 3100MHz.
Q: Are you able to run Prime95?
A: I tried to run P95 blend @ 3.2Ghz, but it could only get to about 1.5 hours and then returned an error.
Q: What is the highest frequency you have used to successfully complete SuperPi 1M?
A: 3306.1MHz (12x275.5MHz).
Q: What is the highest frequency you have realized with the 3800+ Venice (suicide screenshot)?
A: 3419.2MHz (11x310.8MHz) @ 1.664V.
ALL OF s7e9h3n's SCREENSHOTS
Post #1,Post #3,Post #7,Post #12,Post #13,Post #28,Post #42,Post #51,Post #58,Post #98,Post #195,Post #245,Post #251,Post #274,Post #276,Post #282,Post #289,Post #296,Post #426,Post #433,Post #435,Post #450
WOW Conrad... You're going to be my favorite moderator :) Just WOW :thumbsup:
EDIT: I'm such a dork :lol:
I have a winchester that will boot and run all day at -40c :D and im not joking.
Quote:
Originally Posted by s7e9h3n
That must be a sticky ;)
If I'm truly your favorite, my Paypal account is now accepting transfers in US funds. I accept anything in $100 increments. :D As far as making this thread a sticky, there really isn't a point in doing so at this time considering this topic is always on top. We'll see.
A big Thank You to s7e9h3n for making this thread possible. If it wasn't for you, our speculations regarding Venice would lead to heated debates. :toast:
No, Thank You Conrad, you've made it possible now for me to navigate my own thread. :D
i think we need to take down some stickies before we add more, otherwise there will be a whole page dedicated to just stickies and that would suck big time.
if you can run the 3500 on air and just OC as you did with 3800 would be ver nice to see how far can get a lower cpu and as i said before thx for all the screen shots and all the info you have posted here.....
i wanna dfi ultra-d too :P
Unless you have cherry modules these venice cores are looking to have a very bright future. :banana4:
Many people have asked him to test his 3500+ too and he hasnt, I think he made it clear he does not want to because he is planning to sell it and leaving it unopened in the box would get him more $ but he needs to sell soon before everyone buys one on zipzoom or newegg. He seems like a very busy man and was bothered enough to test the 3800+, thanks for doing that. We will see alot more venice testing in the days to come
My thoughts exactly :slobber: great cpu so it seems :thumbsup:Quote:
Originally Posted by clockedOut
btw why do you get 12260 MIPS @ Sandra Dhrystone CPU Arithmetic benchmark @ 2.64Ghz?
I don't even get near that, even though my Winnie 3000+ runs @ 2.6Ghz.. I get like 11140
Thanks a lot s7e9h3n! But we all really hope you'll return soon with some U-D 4x256 results (even w/out dc), however we respect if you don't (traitor :P).
Cheers :toast:
Not really useful above 3.1Ghz then IMO :)Quote:
I tried to run P95 blend @ 3.2Ghz, but it could only get to about 1.5 hours and then returned an error.
If I recall this correctly, they fixed a CPU errata, so with less errors to correct the CPU probably does the arithmetic faster.Quote:
Originally Posted by Garrett
still usefull enough :D
:D THX a lot for You wonderful work Stephen, especially for the FAQs - will help a lot of poeple in future I think.
Also THX to You Conrad for Your help.
Noticed the 3,4 shot and it's very impressive, this is real FX-55 performance and nobody now can still imagine what a 4000+, a 4200+ or even the FX-57 is able to do...........
So I'm looking forward and hope You'll soon get one of these baybies - I'm really nervous to see the first screenie of a FX-57 on single phase.......
:toast:
It's kind of disappointing the small increase from phase change cooling these new 90nm parts seem to exhibit. I mean ~2.8-2.9GHz stable on air and ~3.1-3.2GHz on phase? For some it might be worth it, but I'm a cheap bastard. :p:
THANK YOU ONCE AGAIN FOR ALL THE INFO, s7e9h3n! :)
Thanks s7e9h3n for all the screens, btw is the 3500+ for sale ? :D
Excellent Info/Review test!
Thank you
Thanks for your time , i want one of those chips :D.
I can´t wait
wow ! great oc !
Nice FAQ :banana:
Great job on FAQ Conrad. Very time consuming isn't it:D.
Obviously its a nice chip, what we really need now is tests with 4x sticks of ram. 4x 256M and 4x 512M and see if the memory controller has been improved significantly.
See if you can borrow 4x 512M of decent ream from someone for testing.
Same here but the Venice is damn impressive for sure.Quote:
Originally Posted by RAndomaN
Thanks. I had to review the whole thread anyway, so it wasn't a big deal. It took me about an hour. :cool:Quote:
Originally Posted by skate2snow
For those inquiring if the 3800+ is for sale, I believe longshot is the new owner.
lol, u rock conrad as do you s7e9h3n for having amd people who love u and doing these tests... :banana: :toast: can't wait to get my hands on the venice :slobber:Quote:
Originally Posted by s7e9h3n
Not true.Quote:
Originally Posted by HARDCORECLOCKER
http://news.com.com/AMD+sneaks+strai...l?tag=nefd.top
Read the FAQ in the first post :slap: ;)Quote:
Originally Posted by Steven
http://www.xtremesystems.org/forums/...579#post762579
EDIT: Stephen, this has been a very interesting thread. Thanks for your work :bows:
3419@310, very nice, hope ya get that dfi board
speakin of which, had to pick me up a dfi lp nf2 ultra-b myself today, fried my nf7-s 2.0 last night checking the voltage on vdd w/ multimeter.....
Europe is scheduled for an April 4 release.Quote:
Originally Posted by Steven
Welcome to XS Hector - love the name! :welcome:
April will be a great time in the AMD camp :up:
this guy's name seems familiar, wondering if its the same person. If it is, where does he find the time to post in forums?Quote:
Originally Posted by Hector Ruiz
Somehow I doubt the CEO of AMD would take the time to visit our humble site, and even then I doubt it would be with his real name . . .
exactly what I was going to say. We get reps from ATI, DFI, G. Skill, OCZ, etc,but as far as I know we don't get direct contact with any AMD reps on these forums. Although it would be impressive, it would certainly be very odd for Mr. Ruiz(CEO of AMD as stated above) to visit us. So no, its not him.Quote:
Originally Posted by matt9669
Heres a link to Ruiz's bio page on the AMd site for anyone that wants to see it. http://www.amd.com/us-en/Corporate/A..._11569,00.html