A different way to test memory bandwith....

[Jess1313]

now adays you see every one using sandra to test memory bandwith...dont get me wrong, it does a good job at it, but check out this program i have, i found it back when i was messing with k6-3's and stuff. it usually showes ram bandwith slower that sandra, but is more responsive to tweaks, is a lot faster to run, and gives you a graphical display of read(green i think) and write(red i think). give her a try and tell us what you scored.(BTW DO NOT SET PRIORITY TO HIGH, YOU WILL CRASH ;) )

[Jess1313]

heres a screenie of my nf7@244fsb@3-4-4-8-2t timing, not so good, give me a min or so and ill get one of my shuttle in my sig posted...

BTW=sometimes CPU clock isnt accurate...

[Jess1313]

my shuttle :cool:

[falcx]

Heres what mine is at.......I kicked my fsb back to stock and have my timings at 2.5-3-3-6. Rig is in sig. Running 2 sets of 2x256 Kingston Value Ram DDR400.

[Drec]

weeee my cheap ass ram

[Jess1313]

thanks guys. wow the p4 seems really smooth compared to the Athlons...

keep em coming.:)

[disk11]

Specs since I haven't updated my sig for some time, all hardware the same:
215*8, 2.5-2-2-11

[Jess1313]

amazing how similar the barton curves are

thanks for testing guys...

[FTC]

Something wrong with laptops... my centrino 1.5Mhz w/DDR2100 CAS2 mem gives averages of 60Mb/sec with this tool (expected results and confirmed with Sandra should be around 2000Mb/s)

Regards
FTC

[Tedinde]

Tried it on my AMD64 rig. I get the same scores weather im @ 200fsb or 250fsb??

[Jess1313]

Something wrong with laptops... my centrino 1.5Mhz w/DDR2100 CAS2 mem gives averages of 60Mb/sec with this tool (expected results and confirmed with Sandra should be around 2000Mb/s)

Regards
FTC


humm im not sure there i have used it on about 11 or so comps. usually if sandra is about 2000, it shows about 1,500mbs. but it usually responds well to adjustment.

[Jess1313]

Tried it on my AMD64 rig. I get the same scores weather im @ 200fsb or 250fsb??


well is you mem bus, running the same as your htt? or are you running a ratio, if you are do you use the same mutltipier at 200 as you did 250?. a64's are really weird, sometimes the multiplier can afect ram mhz.

look in cpuz under the memory tab. under ratio, sometimes it shows it being, for exaple cpu/12. or cpu/14 or such as that. with "cpu" meaning "cpu multipier then it it was cpu/12, and you had a 10 multuplier then you would have a 10/12 (5/6) ram deviser, then if you changed to an 8 multiplier then it would reduce the ratio to 8/12 (2/3) effectively lowering your ram mhz.

:D
p.s.=Tedinde, your the one with the k8n neo PL over at xtreme arnt you?
check and see what its are and report back

thanks

[Tedinde]

Same multiplier 1:1 250fsb or 1:1 200fsb I just tested @ 280fsb 1:1 2800mhz Sisoft scores go up but this stays the same

[Jess1313]

I thought this to be very strange but I tried what you said & you where right.

I figured it out though & you are using the refresh buton & it has to be closed & then reopened. Strange though.

It doesnt do this when changing mem timings just when changing FSB.

[Sjaak]

Hmz....anyone else with a P4 to compare? running my RAM 5:4 212 Mhz

[Jess1313]

Youll notice the #s on the left are usally close to what sandra reads.

I use the prog for setting up because its quicker than sandra & then after I get where I need to be I run a sandra.

That P4 score is insane :attn: .

[Jess1313]

bump. :p

[Yamiyanazz]

2500+ @ 200x11
Asus A7N8X rev.2
OCZ PC3200 256MB x 2 matched set

[DarkMatter13]

Here are my scores in this Bench, and Sandra, respectively:

System Specs
Pentium IV-E 3.0GHz @ 3.75GHz (250X15)
5:4 Corsair XMS PC3200 @200MHz 2-3-3-6 1T

Sandra Results (http://www.ocforums.com/attachment.php?attachmentid=31602&stc=1)

[enduro]

Here's a pic of the computer at AUM's computer lab's bench. I can't believe I'm using this computer. It's a Gateway P4 1.6 Ghz w/ 128 Mb of SDRAM

[Jess1313]

as usualy intel is pwning rambandwith..... but oh wait, let me bench my KV8 pro......hummmmmm........ :attn: on guys nice scores

[deathstar13]

my 2.8E with 2x512 kvr bt-d43 287fsb 5:4 2.5-4-4-7 230fsb mem

this is relaxed and i just did it without trying for a high core. ill do that next :)

yeh some days i take it easy and run only 4ghz you cought me :P

[deathstar13]

and one fully tweaked.
4112 mhz 293fsb 5:4 2.5-4-3-5 235fsb mem

i do like it as its alot faster than sandra,long as it shows to be as accurate and more reliable.

[hey560]

Heres mine, Ram is Samsung PC3200 2x512@200Mhz, 3,3,3,11 (i know weak timings trying to better them)

[Jess1313]

its great for setting up because it is quick. The #s to the left will be real close to sandra & it will respond to ajustment well. Its best to close it & reopen bettween ajustments but even with this its faster than sandra & after tweaking you can run a sandra with it being thats what everbady is going buy.

[The Wicker Man]

Mushkin Level 1 3500 218 FSB 2-3-3-11
Hopefully I can get a little more bus with NF7-s v2, should be here this weekend...... :santa: :santa:

[theELVISCERATOR]

everyday settings....

[aNTiChRisT]

Rig#1 averaged 2.5gigs and peaked at 5gigs, with all programs running.

Nice little program.

~t0m

[bobad]

Cool little app! Thanks!

[sx7r]

Here's mine

[Jess1313]

Here's mine


WOW sx!!

i just noticed this, what sort of rig do you have??

amazed
-Jess-

[sx7r]

WOW sx!!

i just noticed this, what sort of rig do you have??

amazed
-Jess-

3.2E @ 3840MHz, IC7-G
240FSB 1:1, OCZ 3200 EB
XP-90, NB1-C

[larva]

I'm not sure what this is supposed to be telling us, but here's mine.

I assume the X axis is block size. And I'm assuming the upper curve is read and the lower is write.

Also, what's with the SSE, MMX, and normal? They seem to change randomly as you run the test repeatedly, but the results are the same.