Screenies of Platnium EB -

[ryanpgroovy]

I have received several comments about how can EB series @ 2.5-2-3 compete with other brands of memory 2-2-3 modules , let alone OCZ Platinum series 2-2-3.

I am on my way to germany tomorow , but it bothered me so much that people did not understand what we are trying to say
in our documentation (http://www.ocztechnology.com/displaypage.php?name=Enhanced%20Bandwidth%20Techno logy)

So I managed to run off some quick benchmarks of our new platnium EB before I go . These are buffered sandra and I know from experience they vary from board to board.

Here are what the scores reveal

DDR 400 at 2-3-3-5 ( OCZ platinum 3200 2x512 )

http://www.ocztechnology.com/images/motherboards/eb0.jpg

DDR 400 at 2-2-3-5 ( OCZ platinum 3200 2x512 )

http://www.ocztechnology.com/images/motherboards/eb1.jpg


The EB platnium 3200 2x512 running at 2.5-2-3-5

Not only does this beat normal CL 2-3-3 products but in addition beats our own platnium 3200.

http://www.ocztechnology.com/images/motherboards/eb3.jpg


also to show off I ran the EB at 500@ 3-3-3-5 vs 3-2-3-5



http://www.ocztechnology.com/images/motherboards/eb5.jpg


and


http://www.ocztechnology.com/images/motherboards/eb7.jpg



now I only had a few minutes and one pair to work with but this is a sneak peak at what we are trying to do and at some of the basic things we have acheived

[ryanpgroovy]

Also they are available in 2gb kits ( 1024x2) but they dont seem to OC as well check it out if you like (http://www.cluboc.net/reviews/memory/ocz/3200_platinum_2gb/index.htm)

not like thats unexpected

[unregistered]

This confirms pretty much the results we have been gathering for the last 3 years, starting with this article:

http://www.lostcircuits.com/memory/ddr2/3.shtml

where se showed that there was no impact of CAS latency on performance. The new boards and speed grades offer better granularity, therefore, small differences are becoming visible now that were swept under the carpet with the slower system and the older SiSoft benches but in general the message is very clear.

[OPPAINTER]

Thats nice but I would bet that it would slow down your FPS in 3D no matter what Sandra thinks.

OPP

[ryanpgroovy]

Opp,

How much would you bet :-)

Just for the record , Were not actualy claiming its better than 2-2-3 , just that its comparable.

[OPPAINTER]

Is this just different memory,,, because if I take the mem I'm useing know and run your timings against my 2-2-2-5 I know my timings will win in 3D.:)

OPP

[ryanpgroovy]

Were doing some diferent things , we published a huge doc on it , applied for a patent ect , and they are new dies .

http://www.ocztechnology.com/displaypage.php?name=Enhanced%20Bandwidth%20Techno logy

Also there's no doubt that 2-2-2 would run faster , We are aiming at reducing the already small effect of dropping cas latency. Thereby making these comparable to 2-2-3 products.

In Addition we want to provide good overhead at low voltage :-)

[OPPAINTER]

I see,, very good.

OPP

[IvanAndreevich]

ryanpgroovy
So.. internal timings of these chips are reduced? Never use Sandra! Try something along the lines of CacheBurst32. Find it via google ;) It will let you test latency as well which will be very important in things like games (Lobby High loves low latency, for example).

[ryanpgroovy]

Ah , I might have forgeten to mention , we are also with the help of some additional technology we have been working on to produce these in 1gb modules

for those who want low latency performance and high density

[OPPAINTER]

BTW:

I'm writing a motherboard review right now, I'll be showing off some of your Cas 2.5 3700 mem in it. I needed something that would run at the mobos limit of 2.78V that could do 250Mhz:D
Your 3700 was the ticket.

OPP

[ryanpgroovy]

Ivan ,

I understand , I am actually on my way to germany tomorow and only had a few minutes to run a quick test.

[ryanpgroovy]

Opp ,

Groovy. The gold is good for those mild low voltage overclocks.

[STEvil]

I like the direction this is going ;)

Is this on the market now? I may have to sell off all my extra ram and get a pair of 512's ;)

[ryanpgroovy]

It starts shipping I think tomorow.

I am hoping that all you overclockers can hit 500 or so at 2.5-3-2 . You guys know I test in a controlled environment

[unregistered]

Opp, you would see some 0.5% performance hit at worst, I had this discussion going on with Tarinder at Hexus and those were his worst case scenario numbers but ... I am listening to everyone who has some real numbers or data. Suffice it to say that in the extreme case 0.5% can mean a lot, like in a 3DMark2001SE score of 30,000... hehe

[OPPAINTER]

Originally posted by unregistered
Suffice it to say that in the extreme case 0.5% can mean a lot, like in a 3DMark2001SE score of 30,000... hehe

Yea it gets even worse at 32k :D

OPP

[ryanpgroovy]

Well with the additional overhead were getting from the EB ( compared to older bh-5 based modules )

I would hope a madman overclocker like yourself you run the fsb higher :-p

[OPPAINTER]

This is the 3rd highest 3Dmark score in the world. I ran this with my mem doing 272Mhz, agressive timings. This score was all memory bandwidth powered.

OPP

[ryanpgroovy]

hehe

I just wonder what you could do with some Eb , being that i never got any Bh-5 to run above 480 , and I can get these to 500 2.5-3-2

[OPPAINTER]

Originally posted by ryanpgroovy
hehe

I just wonder what you could do with some Eb , being that i never got any Bh-5 to run above 480 , and I can get these to 500 2.5-3-2

I'm getting a unlocked P4EE chip this week and will be back to good old Asus P4C-800 and Abit IC-7.
Good boards for testing your Eb.
Send it over:D

OPP

[eva2000]

nice stuff Ryan, surprised in the very little difference between ras precharge of 3 and 2 in the DDR500 benchies

[blinky]

do these chips contain CH5, or did they go with hynix or samsung??

[ryanpgroovy]

Eva they are pretty early benches , so we will see what people come up with

[STEvil]

What kind of volts were needed for DDR500 3-3-3-5?

[ryanpgroovy]

2.8