• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Anyone test 4x512 with the Rev E's?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
K

Kil4Thril

Member
​
Joined
Apr 29, 2005
Location
Charlestown, IN
As stated, has anyone tried this to see if the memory controller implementation is truly improved? I want to use my full 2 Gigs of XMS4400, but I'm limited to awful speeds by my 4000+ :bang head
 
Guys, it's not like 2T is even that big of a deal. It doesn't hurt as much as everyone claims, except for in Sandra...it was worth a smidge over 1% in 3D benchmarks, and about a second in PI. Not an earth-shattering difference. There's absolutely no doubt that 2 GB at 2T would be a lot nicer in daily usage than 1 GB 1T.
 
Gautam said:
Guys, it's not like 2T is even that big of a deal. It doesn't hurt as much as everyone claims, except for in Sandra...it was worth a smidge over 1% in 3D benchmarks, and about a second in PI. Not an earth-shattering difference. There's absolutely no doubt that 2 GB at 2T would be a lot nicer in daily usage than 1 GB 1T.
Click to expand...

Interesting..... I'd love to see some data to that effect.:)
 
1T is only a problem when using 4x 512 right? 2x 512 and 1T are still good with Rev E cores right?
 
indeed..... my venice will be here Tuesday or wednesday so I'll do the whole 2Gig 2T vs 1Gig 1T comparison......

I'll do the comparision with my 4 sticks of TCCD..... right now I'm sitting on 293 2.5-4-3-7 1T. Hopefully I can get pretty high with all 4 sticks with 2T.
 
That's what worries me-having to go to 2T AND dropping my HTT down. 2T isn't scary, but I want speed. I've been told numerous times that my 4000+ should be able to go with 4 sticks due to an "upgraded memory controller", but I could never even get into the BIOS with them, but I can run any 2 of the 4 in dual channel up to 315 FSB in MemTest. Winblows is another story- but I'm working on it.

BTW- who would be the networking guru around these parts as I have a problem.
 
Quick SuperPI 1M

2754MHz, mem at 250MHz 5-2-2-2, WinXP, no tweaks

1T: 31.637 sec
2T: 32.085 sec

Less than half a second difference. And just to put it into perspective, 2714 MHz, mem at 247MHz 1T gets me 32.094 sec. So basically, 2T takes off a whopping 3 MHz. Not especially earth shattering.

And now something for all the TCCD fanboys to chew on. 2754 MHz, 1T, but lets slacken up to 5-3-3-2. Brings me up to 32.086, i.e. exact same loss as 2T. Funny how no one feels uncomfortable running 7-3-3-2.5 timings and the like (yuck!) but talking about 2T is enough to get most people barfing. Both are worth almost exactly the same amount.

And I know what you're thinking. It's just one benchmark. That it may be, but it certainly tests memory, latency and CPU speed in a rather balanced manner.

After all, Sandra is just one benchmark as well.

These results, in fact tally up exactly to some 3D runs I made comparing the same stuff about a year ago, but I don't have time to do any of that right at the moment.
 
Kil4Thril said:
I've been told numerous times that my 4000+ should be able to go with 4 sticks due to an "upgraded memory controller"
Click to expand...

so that you can run all 4 in PC3200 instead of PC2700.
 
Gautam, I think your point would be better supported with a 16mb or 32mb superpi. And you know how people like pretty pictures..... some 4 stick vs 2 stick pics of a 3D benchmark of your choice would also go a long way..... ;)
 
i seem to be getting a 1 second difference in SuperPi.

see sig for specs, only difference between 35 secs and 36 secs is 1T vs 2T
 
Scrufdog said:
i seem to be getting a 1 second difference in SuperPi.

see sig for specs, only difference between 35 secs and 36 secs is 1T vs 2T
Click to expand...
Try out a version that shows the milliseconds, so you can see the precise difference.

Houserat:
Yeah I know all I showed was very very limited in scope. I'll run some 16M and 32M and 3D in the near future.

I've got at the moment 2x256 and 1x512. Not 4 sticks. But I can put 1T and 2T head to head with what I've got. I ran 2T for months with 2x512, and believe me, there was nothing to be felt from switching down to do 1T. I thought that my synthetic benchmarks would skyrocket immensely, but they didn't.

I remember that I went up from 24033 to 24427 in 3DMark01 from 2T to 1T if that helps...2% difference there.

Those that know me know I'm a bencmark fanatic, and even I'm having trouble finding a significant difference between 1T and 2T. I'm sure that running 270-300 2T would score well in benchmarks if that's what you're looking for, and would also be excellent for gaming or anything else that requires a lot of memory. I wouldn't sweat it.
 
You must log in or register to reply here.

Similar threads

Al_bundy
Alucryl - 3DFX - CNC - Retro pc
Replies
2
Views
388
EarthDog
EarthDog
W
Not enough memory when scroll the Facebook
Replies
2
Views
525
The_Jizzler
The_Jizzler
R
Still Considering New Laptop
Replies
8
Views
860
DeadSmiley
DeadSmiley
don256us
FAQ: What is T32's current standing in the world?
Replies
8
Views
468
dfonda
dfonda
tacitus
Building a silent rack-mounted workstation with the Silverstone RM51
Replies
4
Views
2K
Zerileous
Zerileous
Back