Toledo has weaker mem controller than Manchester

[ben805]

And Why is that? I just bought a X2 4400 to play with, at the exact same clock speed and settings (10 x290, 290 @3-3-3-8) for whatever reason the memtest reported the 4400's memory bandwidth about 600 MB/s less than both the X2 3800 and X2 4200 that I have, with the extra 1MB cache I would think the memory bandwidth would blow the Manchester out of the water, the 4400 also score worst in Everest and Sandra than 3800 and 4200 as well, and the worst part is....this POS 4400 can't clock any higher than either one of the Manchester and generated crap loads of heat!! man I'm so sick of keep buying the 'budget' cpu in hope to hit higher clock, in the past 2 months I'd went through TWO X2 3800, a X2 4200 and now the X2 4400.....dang it I think I better off jump on either the X2 4600 or FX-57.

I'm running two instances of SuperPI 32M now, will try different BIOS and see if things will get better.....this sux!

[Absolute_0]

Haha, people aren't going to believe this easily, you'll have to make a thread similar to your Venice VS San Diego shootout. I've always been a fan of 512kb cache, cheaper, cooler, seem to OC better, and about the same performance.

Anyway, you should document this better with some screenshots. Maybe your 4400+ just has a crappy controller and your other X2s are normal.

[ben805]

I'm not going to do another shootout like I did with the Venice vs SD that's for sure LOL there's always naysayer out there and they can believe in whatever they want, this is the second time I fell into this '1MB cache is superior to 512K' hype....be it single core or dual core, there's definitely no 3rd time a charm for sure All I know is...two time being a sucker is enough to stop all those doubts in my mind, back to this POS 4400 I got...if it can clock slightly higher than all my other Manchester then I wouldn't mind too much about the memory bandwidth lost but hot damn....clock worst plus the extra heat it puts out? do I have the suckiest luck! this sucker is going straight to ebay.....now I have to decide if I'm going for the X2 4600 or FX-57, arrgghhh!!

[Absolute_0]

pick up a week 27 3800+, they seem to clock pretty good

[IvanAndreevich]

Absolute_0
Chips with 512KB L2 cache do not run cooler than chips with 1024KB L2. Cache doesn't contribute to the heat dissipation

[s7e9h3n]

There's a reason why I chose to test the 4600's as opposed to the 4400's Larger l2 cache makes OC more difficult.

[Major]

My X2 4600+ had a MUCH better memory controller than the 4400+ I have. Not sure why ? you would think it would be the other way around.

[s7e9h3n]

My X2 4600+ had a MUCH better memory controller than the 4400+ I have. Not sure why ? you would think it would be the other way around.

That should be expected Now if it had a better memory controller than say a 4800, then there's something wrong. BTW, the 4600+ I sold OPB has literally the stongest memory controller I've EVER seen on an x2. Nothing I've seen can even come close.....

[ben805]

ok...atleast I now know it wasn't just my imagination regarding the weaker memory controller of my 4400 the stepping I got is:

ADA4400DAA6CD
CCBWE 0532 TPMW

It kinda sucks that I can't even prime at 2.85Ghz under a single stage so I decided to yank out the IHS, the core is much bigger than both 3800 and 4200, looking at the core prints on both side it appears the contact was pretty lousy. So far I've dropped about 10'C but it's still running hotter than all my Manchesters, still running two instances of SuperPI 32M at the moment and the evaporator is at -33'C(uei DT200) and Smartguardian reports cpu at 253'C, I'll have to see how high can I prime it and if it can't do atleast 2.90G then I'll get the 4600 next, 1MB L2 can kiss my arse...

My X2 4600+ had a MUCH better memory controller than the 4400+ I have. Not sure why ? you would think it would be the other way around.

Have you ever try to run your 4400 under the LS? how high could it prime?

[ben805]

Man....this shiat is sad, I'm running 20Mhz higher on the 4400 and the SuperPI 32M still can't compare to my cheap'o 3800

3 days ago: X2 3800 @2.90Ghz



Now: X2 4400 @2.92Ghz



20 Mhz higher, but 30Sec slower....arrgghhh

[Perc]

LOL man this is crazzy! im glad i bought the 4600x2 off of major instead of buying the 4400x2. i was within one day of buying the 4400x2 for the very same reason you did then i seen majors for sale at the same price as the 4400x2 so i grabed it and im starting to think im lucky i did. now reading your thread im damm glad i did! but then again this is just ONE 4400x2 it could be that your just unlucky with that chip who knoes well man i bid you good luck and hope you find what your looking for. nice thread man it will help alot of people in deciding what x2 not to get...

peace perc,

[Absolute_0]

Absolute_0
Chips with 512KB L2 cache do not run cooler than chips with 1024KB L2. Cache doesn't contribute to the heat dissipation

I'm just going by what i hear. Everybody tells me that SD's hotter than Venice and Toledo is hotter than Manchester. And Ben805's experience only strengthens the theory.

[Vapor]

It would make sense that there is a higher heat output of 1MB L2 chips....there're more transistors.

[5days]

All you people saying 512KB caches have better Mem controllers than 1MB, then how come my Venice sucks. .

(Yes I already know the answer, I was just having a rant )

[ben805]

My temporary spare 3000+ Venice that I bought from Microcenter about a week ago (0533 SPWM) has excellent memory controller, no problem running 2GB Ballistix at 297 @3-3-3-8 too bad it only prime stable at 2.80Ghz under single stage, nevertheless it's a good cpu to push memory to the max

[Major]

Have you ever try to run your 4400 under the LS? how high could it prime?

It was prime stable at 3040mhz under the LS, but I've since made mods to the mounting I think it should have done around 3.1ghz with better contact that I have now. My 4600+ did 3180mhz prime/pi32 stable with the improved contact.

[Major]

That should be expected Now if it had a better memory controller than say a 4800, then there's something wrong. BTW, the 4600+ I sold OPB has literally the stongest memory controller I've EVER seen on an x2. Nothing I've seen can even come close.....

what I was saying is I would have thought the Toledo core would have had a beter memory controller vs the Manchester. That seems to be the way things go with single cores San Diego vs Veince

[mongoled]

It would make sense that there is a higher heat output of 1MB L2 chips....there're more transistors.

This is wht I like to see, some BASIC LOGIC applied to a simple situation.

Lets see .......are

two identical cars heavier than 1 car (YES)
two identical buckets of water heavier than 1 bucket of water (YES)
two fit twin sisters in bed with me better than 1 sister (Hell YES!!)

so why cant peeps apply these simple idea to other things

YES, 1mb cache IS going to produce more heat than 512kb cache

doh (insert homer icon here)

mong

[ben805]

It was prime stable at 3040mhz under the LS, but I've since made mods to the mounting I think it should have done around 3.1ghz with better contact that I have now. My 4600+ did 3180mhz prime/pi32 stable with the improved contact.

diammn....you've got a very nice 4400, lucky you!! this POS I got won't even prime at 2.88 left alone 3.0Ghz core 1 error out much quicker than core 0 even with the IHS off. For now I'm running it at 2.80 till I get a replacement...sigh. I really don't understand how and why some people insist to fool themselves into thinking running these 4400 or 4800 at 2.80Ghz = TWO FX-57

By the way Major, how do you like your FX-57 vs the 4600 you had? I'm still trying to decide which one to jump on....I'm leaning toward the FX

[s7e9h3n]

All this talk of the 4600's makes me miss my two gems:
OPB's 4600 (which I've had at 3300 @ 1.488Vcore) and of course is THE 10K PcMark04 cpu

MikeGuava's 4600 - which was atop PCMark05 for quite awhile - not sure if it still is

[situman]

ok...atleast I now know it wasn't just my imagination regarding the weaker memory controller of my 4400 the stepping I got is:

ADA4400DAA6CD
CCBWE 0532 TPMW

It kinda sucks that I can't even prime at 2.85Ghz under a single stage so I decided to yank out the IHS, the core is much bigger than both 3800 and 4200, looking at the core prints on both side it appears the contact was pretty lousy. So far I've dropped about 10'C but it's still running hotter than all my Manchesters, still running two instances of SuperPI 32M at the moment and the evaporator is at -33'C(uei DT200) and Smartguardian reports cpu at 253'C, I'll have to see how high can I prime it and if it can't do atleast 2.90G then I'll get the 4600 next, 1MB L2 can kiss my arse...



Have you ever try to run your 4400 under the LS? how high could it prime?

well there's ur problem. A week 32 chip. My week 29 TPMW 3800+ kicks ass. Not as good as some, but better than many.

[sauria]

I have a week 29 3800+ X2 E6 that runs 2.7GHz at stock vcore on air.