-

[SupermanCK]

Yeah...
I also read somewhere that DDR should go up to about 500MHz...DDR500...and eventually DDR800...which is Dual channel DDR400....why is technology advancing so slowly...j/k

 

[THUGSROOK]

todays vid cards have 4.5, 4, 3.6, 3, and even 2.5ns memory - why dont systems?

 

[CrawlingEye]

Originally posted by: THUGSROOK
todays vid cards have 4.5, 4, 3.6, 3, and even 2.5ns memory - why dont systems?

Easily answered: JEDEC/Samsung/IBM are all making a joint effort at pushing DDR-II.

 

[THUGSROOK]

Easily answered: JEDEC/Samsung/IBM are all making a joint effort at pushing DDR-II.

ok thats a good excuse. may i ask why DDRII is slated for a 166/333DDR and 200/400DDR introduction than?

i ask again - why isnt system memory on par with vid card memory? (ns ratings)

 

[NanoMem]

Originally posted by: THUGSROOK

Easily answered: JEDEC/Samsung/IBM are all making a joint effort at pushing DDR-II.

ok thats a good excuse. may i ask why DDRII is slated for a 166/333DDR and 200/400DDR introduction than?

i ask again - why isnt system memory on par with vid card memory? (ns ratings)

Let's see 2.5ns translates into 1000/2.5=400MHZ which would sync up with available GPU speeds in the same range and of course the large supporting bandwidth in the AGP bus. I think the main show stopper in transferring these standards over to system DRAM is intel's control over the CPU and chipsets. Can you imagine the 6400MB/s memory bandwidth with 400*x2 for DDR? The required multiplier/ratio using the officially supported 133 FSB would be 1:3 ie. memory clocking 3 times faster than CPU. No wonder Intel has cooled down its RDRAM path in favor of DDR. Chip latencies of only 6 ns used in DDR400 are starting to show their edge over 30 ns RDRAM used in RIMM4200 at the same clock speeds in spite of the latest 32-bit RIMM standard release. They basically transferred the dual channel of the two RDRAM socket bus of the mobo into the new 232 pins of a single DIMM therefore leaving the established record high of RDRAM unchanged. I call this just technology repackaging rather than innovation. I am therefore sticking with DDR ie stirring away from i850E. Btw, although the Asus P4T533 sports specs for RIMM4200, the DRCG embedded in that board is rated for only 533MHZ. I am sharing this with you cuz I noticed that you started to lean towards that board when you ordered the P4B533.

Btw, DDR450 and beyond are feasible without cooling Samsung PC2700 chips.

Peace.

 

[St3v3]

OK, something is wrong ... I can boot 160 FSB @ 1:1 settings, I move to 3:4 settings, I can't even boot @ 141, what else do I need to do?

 

[oldfart]

NanoMem, any idea what the "beyond" bench is from? Those mem scores are sick.
Here is a 3400 mem score is a from a guy @ [ H ] OCP running DDR446 on a 4BDA2+. He also got a 13107 3DMark2K1 score with a Ti4400.

 

[Nighthawk69]

St3v3: Your RAM probably can't handle the higher speed.

 

[NanoMem]

Originally posted by: oldfart
NanoMem, any idea what the "beyond" bench is from? Those mem scores are sick.
Here is a 3400 mem score is a from a guy @ [ H ] OCP running DDR446 on a 4BDA2+. He also got a 13107 3DMark2K1 score with a Ti4400.

I had the same reaction as yours! I scrambled all over my calculator to see what this guy was cooking.
1)I suspect it's a different system and most likely based on an RDRAM clocking at 569MHZ which would not be uncommon since PC1200 RDRAM is readily available.

2)I think this might be an R & D bench, and either they're using DDRII at FSB 142.5 with 1:1 cuz if not, with plain DDR, the RAM would need to be clocked to 285MHZ for a DDR570 based on his 4057 score and an efficiency of 89%.
Peace.

 

[THUGSROOK]

ive been running some tests and i came across this oddity....

using a 64mb AGP Aperture with 3:4 i get a blackscreen in the middle of 3dm2k1.
but if i use a 128 or 256mb AGP Aperture it runs perfectly fine

...might help a few of you guys get to 155fsb 3:4

 

[THUGSROOK]

ive got a stick of Sammy Original 2700 comin.
gonna give it a direct comparison to the Corsair XMS

 

[Nighthawk69]

Well, 155Mhz FSB (1.75v true) with 3:4 is working great--I like it a lot more than 160Mhz FSB (1.85v) at 1:1. But, when I try to run 3:4 at 160Mhz FSB, it can't take it. I don't bluescreen, but I do have errors and programs crashing left and right in Windows. Oh-well.

 

[THUGSROOK]

i just verified that Turbo1/2 does work in jumpered mode.

 

[Nighthawk69]

Sweet Thugs! Thanks for the heads-up!

 

[Mage]

NightHawk: Did you use a voltmod to get to 1.75V? How risky is it? Someone once posted on Hardforum that they fried their 1.6 that way

 

[THUGSROOK]

yes we are using the 1.7v wire mod.

 

[Mage]

Hrm, my cpu can't seem to hit 2.4 even with 1.700 Vcore, and if I could reach 150 FSB I'd be real happy... I -was- running at 150 FSB when I didn't have the Live and another hardrive in there, so maybe it's the 300W Antec PSU that isn't cutting it? But it's always ~11.95 or so, and my system only consists of: 2 hardrives, 1 cdrw, 1 sblive, 1 NIC, and 4 low-speed casefans.

So anyhow, if I did a voltmod to 1.750V or something, and could reach 150 FSB stable, I'd be really, really happy. Just... the idea of taking the HSF off again, and possible smudging the arctic silver around, and not knowing how risky the mod is giving me butterflies

 

[THUGSROOK]

the 1.7v wire mod is giving us 1.74-1.76v in windows. thats more than enough for my system to run 160fsb 100% stable in jumper mode.

 

[marylandnorthwoodclock]

Originally posted by: THUGSROOK
the 1.7v wire mod is giving us 1.74-1.76v in windows. thats enough for my system to run 160fsb 100% stable in jumper mode.

Was wondering how this might go if you flashed the P4B533 bios onto the P4B266 as I did with mine. Can't go back to any P4B266 bios of course, but since I've donated my board and P4 1.6a to the local Junior High after doing so..(it works flawlessly and they won't know till they start fooling with the bios, bright kids that they are) anyway, too late for me to experiment with that now. Was just wondering if anyone tried yet?

Regards

 

[THUGSROOK]

cant.
the 533 board uses 512k bios - the 266 board uses a 256k bios.

 

[marylandnorthwoodclock]

so true, but I let a locale "engineering enthusiast" do something...and it could very well be that the Junir High might be disappointed at some future date, in which case, I'll give them another PC I have around.

 

[THUGSROOK]

 

[THUGSROOK]

 

[SupermanCK]

I can run DDR426 with vmem 2.5v and by spd...doh
but i can't complete 3dmark2k1...it closes in the middle of benching back to desktop...
can that be the mem???

 

[SupermanCK]

I can run DDR446.8 with vmem at 2.5v
i had it ran a few time...got 12800+ 3dmark2k1 before i start to mess around with video o/c setting...
then it started to close to desktop...it won't finish benching...i was trying to get it up to 13000...but then the crashes happened...what could be wrong?