Zroc, why MSI 865 PE board? Why not the Asus P4P800? Inquiring mind wants to know!:D
Printable View
Zroc, why MSI 865 PE board? Why not the Asus P4P800? Inquiring mind wants to know!:D
Hehe, Ace, look when I wrote this piece...the first of May ;)Quote:
Originally posted by Ace-a-Rue
Zroc, why MSI 865 PE board? Why not the Asus P4P800? Inquiring mind wants to know!:D
If you remember, Springdale wasn't even out yet...it launched May 21st, officially. The MSI board was out late April 'for evaluation purposes', and was the only 865PE out. When I was doing the work for this, it was just 875P vs. nForce...then I saw that board, and decided to include it. Pretty neat at the time.
If I was to do another roundup today, I'd have used the IS7, actually. And all the results would have been higher (IC7's now score 4900-5000's in Sandras since BIOS 1.4, as does the IS7 with F1 acceleration). The basic conclusion today would have been an 865PE with acceleration is equal to an 875P at stock 1:1, but acceleration (or any PAT trick) does not work at 5:4 or 3:2, so the 875P would be faster at those ratios with equal settings.
But of course, it'll change again next month, because Intel is already shipping the new revision 865PE's that can't be 'tricked' into thinking there's a 533 proc installed when it's an 800...so by next month, it'll be back to how it was originally.
All very facinating ;)
That said, I'll probably just wait for Prescott/Hammer for the next roundup. This piece served it's purpose well when it was written, and everything was all new, shiny and unknown :toast:
No doubt!:) Your piece was like a symphony.:) Bravo! I have to admit I didn't check the date you started this thread but I did buy the Asus SD m/b in early June (12th). It is a very good board especially since Asus broke the code for PAT, or in Asus's words "MAM": Memory Acceleration Mode.Quote:
This piece served it's purpose well when it was written, and everything was all new, shiny and
I am very close to your 875 m/b which was run at 271 fsb and I am running (3G) at 234 fsb, 1:1. When I say very close, I am talking about Sandra Memory bench. I know, I know, you don't cotton to that benchmark.:) I am at 5688@234 FSB. To get there, I am using a Vmem mod. I really can't pump anymore (max 240 fsb at 5:4 or 3:2) out of the cpu even with the most relaxed memory timings. But I am amazed at the board for allowing me to do 234 fsb at 1:1 using Twinmos Winbond BH-5 chips and then, haul arse:D.
Hey Ace-a-Rue I swear I've seen your name somewhere else?Quote:
Originally posted by Ace-a-Rue
No doubt!:)...I am at 5688@234 FSB. To get there, I am using a Vmem mod. I really can't pump anymore (max 240 fsb at 5:4 or 3:2) out of the cpu even with the most relaxed memory timings. But I am amazed at the board for allowing me to do 234 fsb at 1:1 using Twinmos Winbond BH-5 chips and then, haul arse:D.
Anyway I wrote an article on "Pseudo-PAT" otherwise known as "GAT", "MAM", "Hyperpath", etc. My take on it is thus; the ceiling will be NB related at 300FSB no matter the timings, voltage, after-market cooling, or CPU (of course all things being equal their optimized for 300FSB). I belive in 99% cases at or above 300FSB it's the NB which holds one back.
As you know there are no "865's" their actually 875's which failed a number of QC and/or speed-bin. Of course the entire pseudo-pat concept is based on the theory not all 875's were actually tested, and the 865 motherboard owner MAY end up with a chip which is PAT capable.
Here's the rub; I beleive they (Intel) HAVE in fact tested every chip, how else can they then eliminate any sub-par 875's from ending up in a 875 mobo? From what I've seen NO 865 mobo owner has been able to come close to the level of performance offered fom these boards. And in fact, in the IS7 case, once the shipping BIOS is replaced with the GAT BIOS if one doesn't Disable CPC (Command Per Clock) they may not be able to use their memory.
I beleive it was a mistake given the conditions (predilication 865 boards can not, prima facie maintain PAT features) to "sell" end-users on the concept they were getting PAT for less. I'd almost gaurantee, you can your system to 300FSB, at 1.550V, Ratio 3:2 Timings 2.5-8-4-4 all GAT settings on Auto except CPC which is Disable.
You must of been hearing people swearing at me!!:DQuote:
Originally posted by Liquid3D
Hey Ace-a-Rue I swear I've seen your name somewhere else?
Have you visited the Intel assembly line or talked to an Intel engineer about your theory? I don't believe they spend very much time on testing. Probably one very simple trace check for resistance. To much, above a certain number, the chip is moved to an 865 board. Probably mine was so close to making it to a 875 board so I reaped the benefits. I can give my good buddy a run for the money, as the saying goes:), with his Asus CW board. I am the fortunate one. It is certainly worth the gamble to buy a SD board. If it doesn't work out, then RMA it for a CW. Certainly saving $60-70 is worth a try. But, you have to work with a quality vendor like Newegg to get that flexibility.Quote:
As you know there are no "865's" their actually 875's which failed a number of QC and/or speed-bin. Of course the entire pseudo-pat concept is based on the theory not all 875's were actually tested, and the 865 motherboard owner MAY end up with a chip which is PAT capable.
I have....lucky me!:)Quote:
From what I've seen NO 865 mobo owner has been able to come close to the level of performance offered fom these boards.
I doubt very much 300 fsb on air cooling! If I had a 2.4G (12X) in my platform instead of a 3G (15X), I would probably get very close to 270 fsb. But, you ain't going to do it with a majority of 3G's aircooling it even with a great HSF for the cpu and active aircooling for the NB!Quote:
I'd almost gaurantee, you can your system to 300FSB, at 1.550V, Ratio 3:2 Timings 2.5-8-4-4 all GAT settings on Auto except CPC which is Disable.
No I haven't visited the Intel test lines, nor spoke with an engineer, but I hope to this evening, have you? That's the essence of a theory. At some point it makes an inductive-leap, or deductive-claim, based on premesis. So far, you are one of the "Lucky ones." There-in lay the quandry.
I'd agree any tests would have to be cursory, however; wouldn't they be more latency relative, then simple resistance?
I did think your chip had a higher multi, but I wasn't sure which model it was. Just out of curiosity, have you exhausted all your options in attempting to reach a higher clock speed? I can see why you don't feel it's the NB holding you back. I beleive it was the higher speed CPU owners 865 motherboard makers had in mind enabling pseudo-PAT.
I guess one could choose the RMA route but on what grounds would one request an RMA? I'm not implying you would or did, but are you suggesting Newegg would exchange a mobo, for another model, because it's NB wasn't up to the PAT latencies?
I wouldn't feel comfortable fabricating. I feel dishonesty perpetuates a dishonest society. Where it effects everyone is in it's cost. Perhaps the 875 boards would only cost $125 had mobo makers not have to absorb the cost of all the dishonest merchandise returns. In fact it's that same dishonesty which prevents you from changing the multiplier on your Intel P4. In Intel's case they wouldn't absorb the cost of the unscrupulous, and therefore locked out multipliers to those resellers (or whomever) chose to remark CPU's and sell them as higher-costing models. Dishonesty albeit minimized in the seemingly "white-lie" RMA, or in counterfeiting on a large scale know's no boundries, exacting a toll on society which is cumulative. (STep off Soap-box) :)
Nope! So, both of us are shooting from the hip in our theories.;)Quote:
Originally posted by Liquid3D
No I haven't visited the Intel test lines, nor spoke with an engineer, but I hope to this evening, have you?
You should find out by tonight. LMK the results of your conversation.Quote:
I'd agree any tests would have to be cursory, however; wouldn't they be more latency relative, then simple resistance?
I have tried everything to get above 240 fsb.Quote:
I did think your chip had a higher multi, but I wasn't sure which model it was. Just out of curiosity, have you exhausted all your options in attempting to reach a higher clock speed? I can see why you don't feel it's the NB holding you back. I beleive it was the higher speed CPU owners 865 motherboard makers had in mind enabling pseudo-PAT.
Oh please, stop with the preaching! Keep those comments to yourself....Quote:
are you suggesting Newegg would exchange a mobo, for another model, because it's NB wasn't up to the PAT latencies?
I wouldn't feel comfortable fabricating. I feel dishonesty perpetuates a dishonest society. Where it effects everyone is in it's cost. Perhaps the 875 boards would only cost $125 had mobo makers not have to absorb the cost of all the dishonest merchandise returns. In fact it's that same dishonesty which prevents you from changing the multiplier on your Intel P4. In Intel's case they wouldn't absorb the cost of the unscrupulous, and therefore locked out multipliers to those resellers (or whomever) chose to remark CPU's and sell them as higher-costing models. Dishonesty albeit minimized in the seemingly "white-lie" RMA, or in counterfeiting on a large scale know's no boundries, exacting a toll on society which is cumulative. (STep off Soap-box) :)
If you don't know Newegg's policies, how can you sermonize that lying is going to happen. They clearly mark their products that they will accept back for a refund. They say they can charge a 15% restocking fee, but most times they don't exercise that policy. Some m/b's are only for exchange and those are clearly annotated in RED!. They also clearly state you cannot exchange a m/b for a different board. You get the refund, which might take about 7-10 days. Meanwhile you just order the board you want. They are the best on-line retailer around! So, if I didn't like this board I would have returned it for a refund and then ordered a CW board. BUT, the Asus SD board is a keeper!!:)
I merely added the "honesty" honestly as a sarcastic retort to your criticality of my originally attempting to help out. I simply treat others as they treat me. After all not all of us could afford to loose the restocking fee, then buy the more expensive mobo. It may be difficult for some of us to comprehend this when you could've purchased the better model in the first place. Something for nothing when you have something is like nothing I'd condone. :) Likewise I'm sure.
Quote:
Originally posted by Liquid3D
I merely added the "honesty" honestly as a sarcastic retort to your criticality of my originally attempting to help out. I simply treat others as they treat me. After all not all of us could afford to loose the restocking fee, then buy the more expensive mobo. It may be difficult for some of us to comprehend this when you could've purchased the better model in the first place. Something for nothing when you have something is like nothing I'd condone. :) Likewise I'm sure.
I think you are over reacting.
I have not thrown the gauntlet down on the floor requiring you to duke it out in a verbal manner! My statements were straight forward and not wrapped with any malicious sarcasm.
I think our discussion is finished. We are wasting server bandwidth with this continuing dialogue about nothing.
Man awesome comparison! very detailed good info. Did seem like u guys worked on pushing that p4 harder than the nforce2 tho to be fair. Many guys running 225+ with nf2 platforms and as your FSB scale charts showed that could add up to quite a diff.
Not like it would catch the faster P4 anyway but woulda made for nice comparison =)
Was a lot closer than i woulda guessed tho!
another interesting observation the AMD's seem to be a bit faster under stress.
Either way awesome review and them p4's tweaked kick some booty.
Finally made it to 3624 MHZ:banana:
P4 2.4C SL6WF Malay:D
302 FSB, 5:4, 2.5-4-4-7,
IS7-E
OCZ PC 4000 Gold dual channel
On air with Swifty mcx 462+ and TT Smartfan II
Sapphire R9800 pro
Absolutely stable imo (8 hours prime95, 3DMark2k1)
Dam that's 24Mhz higher them me on that board!!! I can't get this IC7-G to clock at 300FSB? I'm thinking of reinstalling my IS7-E. It was actually the best board out of three I have, and the other two are 875 boards?
NICE OC brother!!!
Wow nice work Zroc! Man I give that review "TWO THUMBS UP!"
Hold your horses made it to 305FSB for 3666Mhz weeee
gratz, Keith,
mine stopped at 303 FSB, (no vmods, exept vdimm, but not in use so far) can do 302 quite stable, runnin right now at 291 FSB 1000% stable
btw, how did u manage, u hit with IC7 or IS7-E????????????
No volt mods, no nothing I just have a two sweetheart 2.4C's from Malaysia. L310A735, and Q320A287. Both do 3.6GHz at default vcore, and remain cool (36C to 39C).
My IS7-E was a much sweeter baord ythen I'd known.
running 302 FSB, 2 primes, Toast, i prime failed after 6 hours or so, other and toast flawless, so im running at 302 FSB, best i could get out of mobo and Malay so far, dunno exactly which limit is reached, mobo or malay..., but its good enough so far 4 me,;) new psu still not arrived:rolleyes:
I tried Kunaak's idea of raising PCI/AGP to 35/71, to no avail. But I'll keep trying. I can run 302FSB all day as well, and think the simple answer to this, is not so much our processor's, but the NB. Most Canterwood's see anomalies above 300FSB. Isn't it ironic when you think of it. Intel designs and build to exacting spec's. You have to give their engineers credit. The hihest model the 3.2C, would natrually have a an air-limit of about 3.6GHz higher because of it's multiplier. They knew where the cailing was (concevably) and adhered to it. No more no less. I'm sure had they wanted tyo squeeze more out of Granite Bay (yes I said Granite Bay) they could have built Canterwood from the gorund up. But that wouldn't be as cost saving as redesigning Granite Bay which they already had Fab production in place for. Considering the cost of $2.2-billion, it was a brilliant bit of engineering prowess to convert GB to run strictly DDR and increase it's through-put.
This comparison didn't help me much. I mean it's good but there are too many real world variables that are not covered enough. I have an AMD 2800 (running at 178 fsb=2.4 gig) and I can already hit 17200 in 3DK3. I have actually been thinking about moving to a P4 around 3.2 gig (lower cpu overclocked to around that). Now according to this comparision review I would be wasting my time because my system can already do what is stated. 17000 is nothing, I want something over 20 - 22,000. I want this without any water cooling. Now when something comes along that can do that, then I will upgrade because otherwise it just wouldn't be worth it.
I have a 3.0 o/c'd to 3.5 at 234 fsb. I too have a 9800 Pro and can get 20K+, safely 20.6 and stretching it to 21K+. Food for thought.Quote:
Originally posted by Mike89
This comparison didn't help me much. I mean it's good but there are too many real world variables that are not covered enough. I have an AMD 2800 (running at 178 fsb=2.4 gig) and I can already hit 17200 in 3DK3. I have actually been thinking about moving to a P4 around 3.2 gig (lower cpu overclocked to around that). Now according to this comparision review I would be wasting my time because my system can already do what is stated. 17000 is nothing, I want something over 20 - 22,000. I want this without any water cooling. Now when something comes along that can do that, then I will upgrade because otherwise it just wouldn't be worth it.
That's pretty good. I wish I could sit at your computer and play some games. Where do you live, I'll be right over! :p:
I have been thinking about the p4 for quite some time. I'd hate to do it though and not notice any difference. Even with my system, there are times playing games at settings I like when I could really appreciate more power if you know what I mean.
I would say wait for the A64 or FX official release before making a decision. I would not jump until a month after reading the forum post's to see how systems really perform. I am skeptic with professional reviewers. I trust you and others to state the issues....
FX? You mean the Prescott? I don't figure you are talking about Nvidia.
The AMD64 hasn't rung my bell. Their decision to use registered Dimms doesn't appeal to me, especially since I have 4 256 megs sticks of Corsair PC3500 laying around.
I'm sure Intel is going to have an answer to the AMD64. I just don't see AMD dominating Intel (Intel just has too dam many resources).
he means the amd64 fx-51 cpu, not the prescott
@ Mike89
with my rig, when FSB set to 302, im running 100% stable @292 FSB , easy 20917 3DM2K1, Videocard watercooled, sys on AIR
:D