WinTune97 TechPowerUp Forums Benchmark results POST

[Alec§taar]

08/05-06/2006 WINTUNE97 BENCHMARK TEST RESULTS SUBMITTED BY TECHPOWERUP FORUMS MEMBERS:

* ALL GRAPHS/CHARTS FROM EACH TEST ARE BELOW, on the same page (for easier viewing & each person's standings + category in each listed as well (by category of test taken))...

==================================================

A.) WinTune 97 Benchmark program download here:

http://forums.techpowerup.com/attachment.php?attachmentid=3265&d=1154641298

B.) Information needed in email while sending exported resultset zipped my way:

In the text body of the letter to me, include your:

• name (one you use on the forums here as you nick/handle)

• your system's name

• your CPU type

• videocard type (run tests @ 1600x1200 resolution for BEST results + 32-bit TrueColor)

• diskdrive type

• memory type + amount

C.) Keep exported resultfiles' filenames in this form:

YourTechPowerUpForumNickNameYourSystemCPUName.wtx

D.) ZIP & EMAIL AS ATTACHMENT -> YourTechPowerUpForumNickNameYourSystemCPUName.zip

E.) Email results to:

apk4776239@hotmail.com

APK

 

[Alec§taar]

Cpu Integer Results Submitted 08/05-/06/2006

CPU INTEGER SCORE STANDINGS:
======================

CPU DhryStone:
-----------------

• MikeLopez (o/c DualCore Intel Conroe e6300 @ 3ghz) = 21052 MIPS

• APK (AMD DualCore Athlon64 x2 4800+ @ 2.761ghz) = 14814 MIPS

• MikeLopez (DualCore Intel Conroe e6300 @ stock) = 12698 MIPS

• SupraMan (AMD 3200 @ 2.5ghz) = 6704 MIPS

• EMachines (AMD Athlon 64 3400+ Newcastle (Socket 754)) = 6431 MIPS

• ATHLON3500 (AMD 3500) = 5949 MIPS

APK

P.S.=> MIKELOPEZ'S CONROE "O/C'd" result ROCKS OUR WORLDS HERE (but places after mine @ stock mhz/ghz speed)... apk

 

[Alec§taar]

Cpu Floating Point Results Submitted 08/05-/06/2006

CPU FLOATING POINT SCORE STANDINGS:
===========================

CPU WhetStone:
-----------------

• APK (AMD DualCore Athlon64 x2 4800+ @ 2.761ghz) = 4000 FLOPS

• MikeLopez (o/c DualCore Intel Conroe e6300 @ 3ghz) = 4000 FLOPS

• MikeLopez (DualCore Intel Conroe e6300 @ stock) = 2500 FLOPS

• SupraMan (AMD 3200 @ 2.5ghz) = 1610 FLOPS

• EMachines (AMD Athlon 64 3400+ Newcastle (Socket 754)) = 1552 FLOPS

• ATHLON3500 (AMD 3500) = 1434 FLOPS

APK

P.S.=> Mike's CONROE & APK/myself top @ #1... apk

 

[Alec§taar]

2d Video Results Submitted 08/05-/06/2006

2D VIDEO SCORE STANDINGS:
===================

• APK (BFG NVidia 7900 GTX OC) = 1543 MPixels/Sec

• SupraMan (XFX 6800GT) = 941 MPixels/Sec

• MikeLopez (ATI X850XT PE) = 925 MPixels/Sec

• MikeLopez (ATI X850XT PE) = 854 MPixels/Sec

• EMachines (ATI Express 200 (onboard graphics)) = 305 MPixels/Sec

• ATHLON3500 (unknown card) = 47 MPixels/Sec

* QUESTION: ATHLON3500 - What vidcard do you use?

APK

P.S.=> GeForce 7900GTX OC (mine) did very well here (red bar), but this is 2d testing of video as well, keep that in mind... apk

 

[Alec§taar]

Diskdrive Cached Results Submitted 08/05-/06/2006

DISKSPEED CACHED SCORE STANDINGS:
==========================

• SupraMan (Seagate Perpendicular Recording SATA II) :1479 MB/SEC

• APK (PCI-e x4 Promise SuperTrak EX8350 128mb ECC Ram CacheController driven RAID 0 2x 'WD Raptor X' SATA I) : 1292 MB/SEC

• MikeLopez non-oc'd (Seagate Barracuda 7200.10 (Perpendicular Recording) w/16MB Cache) : 931 MB/SEC

• ATHLON3500 (unknown disk) : 825 MB/SEC

• MikeLopez oc'd (Seagate Barracuda 7200.10 (Perpendicular Recording) w/16MB Cache) : 354MB/SEC

• EMachines (Western Digital 200GB 7200rpm w/2MB) : 230MB/SEC

* QUESTION:ATHLON3500: WHAT TYPE OF DISK DO YOU USE?

APK

P.S.=> Note my score?

That is a combination of "SuperCache-II", a superior diskcaching program to the native 2000/XP/Server 2003 diskcache, AND a 128mb ECC RAM caching controller with Intel IO subprocessor on it running a Western Digital "Raptor X" 10,000 rpm 16mb buffered 2 drive RAID 0...

Still, not as fast as a SINGLE Seagate 7200 rpm perpendicular recording drive SupraMan has (incredible read speeds due to data density on platters)!

I've seen the same basic results on HDTach here on 2 tests already vs. those, even w/ my setup (about as good as you'll get imo @ least on SATA I)

Especially/mostly for server patterns & writes due to the 128mb ECC RAM Caching Controller (w/ Intel I-O firmware CPU) being great for bursting out of cache, & 0% system cpu-usage, + the disks give it bigtime fast seek/access due to 10,000 rpm platters & 32mb of buffers on them!

 

[Alec§taar]

Diskdrive Uncached Results Submitted 08/05-/06/2006

DISKSPEED UNCACHED SCORE STANDINGS:
============================

• MikeLopez oc'd (Seagate Barracuda 7200.10 (Perpendicular Recording) w/16MB Cache) : 29 MB/SEC

• MikeLopez non-oc'd (Seagate Barracuda 7200.10 (Perpendicular Recording) w/16MB Cache) : 21 MB/SEC

• ATHLON3500 (unknown disk) : 4.1 MB/SEC

• SupraMan (Seagate Perpendicular Recording SATA II) : 2.8 MB/SEC

• EMachines (Western Digital 200GB 7200rpm w/2MB) : 2.6 MB/SEC

• APK (PCI-e x4 Promise SuperTrak EX8350 128mb ECC Ram CacheController driven RAID 0 2x 'WD Raptor X' SATA I) : 2.2 MB/SEC

* QUESTION: ATHLON3500: WHAT TYPE OF DISK DO YOU USE?

APK

P.S.=> Things change here in the uncached results though, bigtime! First time I am @ the bottom, lol!... Mike I think that's YOU AGAIN, taking the top-spot! apk

 

[Alec§taar]

Memory Tests Results Submitted 08/05-/06/2006

MEMORY READ SCORE STANDINGS:
=======================

• MikeLopez oc'd (CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)) : 6381 MB/SEC

• ATHLON3500 (unknown ram) : 2862 MB/SEC

• SupraMan (2*512MB Corsair Value PC3200) : 581 MB/SEC

• EMachines (1 GB (2x512) Nanya PC3200 DDR SDRAM) : 528 MB/SEC

• APK (Corsair Matched Pair 2x256mb DDR-400 running DDR-333 185/200mhz) : 324 MB/SEC

• MikeLopez oc'd (CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)): 0 MB/SEC

MEMORY WRITE SCORE STANDINGS:
=======================

• APK (Corsair Matched Pair 2x256mb DDR-400 running DDR-333 185/200mhz) : 2118 MB/SEC

• ATHLON3500 (unknown ram) : 1955 MB/SEC

• SupraMan (2*512MB Corsair Value PC3200) : 1945 MB/SEC

• EMachines (1 GB (2x512) Nanya PC3200 DDR SDRAM) : 1829 MB/SEC

• MikeLopez oc'd (CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)) : 0 MB/SEC

• MikeLopez oc'd (CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)) : 0 MB/SEC

MEMORY COPY SCORE STANDINGS:
=======================

• ATHLON3500 (unknown ram) : 4526 MB/SEC

• MikeLopez oc'd (CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)) : 3368 MB/SEC

• SupraMan (2*512MB Corsair Value PC3200) : 2784 MB/SEC

• EMachines (1 GB (2x512) Nanya PC3200 DDR SDRAM) : 2674 MB/SEC

• APK (Corsair Matched Pair 2x256mb DDR-400 running DDR-333 185/200mhz) : 888 MB/SEC

• MikeLopez oc'd (CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)): 538 MB/SEC

* QUESTION: ATHLON3500 - What kind of RAM do you run?

APK

P.S.=> RAM Write portion: I got #1 there!

RAM Read portion: MikeLopez (on his non-overclocked test of his INTEL CONROE e6300) got #1 there!!

RAM Copy portion: Athlon3500 got #1

apk

 

[Alec§taar]

TO TEST PARTICIPANTS:

If you would like my entire imported DB, it is yours as well!

(Has machines from 1992-present in it, & all types nearly in between from forums all over the net in that time (The file is 100kb in size, zipped, & has 100's of results... literally!))

* Your results are ALL included too!

(That is so you can also verify the results & such vs. one another here, AND vs. machines from the past too if you like... you can do your OWN comparing then!)

My email is:

apk4776239@hotmail.com

APK

 

[Alec§taar]

"bump!"



Apk

 

[mikelopez]

My C2D RAM specs:

CORSAIR XMS2 2GB (2 x 1GB) DDR2 800 (PC2 6400)

 

[Alec§taar]

Results For This Test Are All On The First Page!



Apk

 

[Alec§taar]

Got it Mike...

You RAM type & amount is now added to the first page of results!

(NOW, the ONLY person missng anything is ATHLON3500, & we aren't sure if he is cdawall or not anyhow!)

APK

 

[Alec§taar]

I just hit a higher overclock (11x multiplier X 251mhz FSB = 2770.1mhz on CPU, & 198/200mhz on RAM)!

I only used a 2761mhz clock last time, & 185/200mhz RAM efficiency...

AND, Ketxxx is turning me onto ways to open up more memory bandwidth in my "RAM timing-chain"!

So I know I can pull better than what I posted here earlier... MUCH better!

(E.G.-> I ran it again on this new o/c a few minutes ago? I got 2248 mpixels on my video score (& only hit 1543 mpixels on the original test whose settings above you see!))



* If I get some more folks or diff. results from those that have taken this test last round? I will include them in a new set of charts... no hurries though!

APK

 

[cdawall]

nope i am a 3000+ and my system is quite simple viking ddr266 (at near 333) maxtor 60gb ata133 not sure the cache

 

[mikelopez]

Wintune seems a little erratic; check out the cached speed results:

 

[cdawall]

you know i think the c2d may be just alittle past the wintune's programming

 

[Alec§taar]

you know i think the c2d may be just alittle past the wintune's programming

I'd think not personally, because it's typical Ring3/RPL3/UserMode synthesized benchmarks written in Win32 API code & C/C++ language iirc.

Some stuff may be "off" (like you see folks here saying about temp monitoring apps here recently) - it's mhz appears to be a hand-written one, & it's "off" some, & its CPUid functions aren't prepped for the CPU's that came after the time it was written in:

486-PPro-Pentium I era

HOWEVER, the real main point here is?

That a "benchmark-is-a-benchmark" & as long as we all run the same test, with the same data? It's even (in the software conditions!).

The Win32 API is the same for the MOST part, & especially here in the things this test does... it's not asking for NEWER features in it either (there are some, but the bulk of them largely are the same).

Compilers may be a BIT more "optimized" now, but not much more than those in 1997-1998 really... in fact, it's one of the SLOWEST THINGS to have catch up usually, because note (iirc, assuming that)?

They really haven't!

See here:

E.G.-> Even when trying to use SSE2/SSE3 instructions in an app like SuperPi & mods for those? At best, I got 1 second 'shaved off' my time using them, vs. x87 typical fpop processing methods in std. Super-Pi 1.4... not much really!

Going from 33-> 32 -> 31 using them (may be a LOT looked @ in terms of milliseconds/nanoseconds, the timescale in which CPU's live) but it's not a lot to cut off ONLY 1 second imo, where WE are - the real world.

APK

P.S.=> Some compilers also optimize better & for BETTER-NESS in specific areas...

A good example of it is the SETI@Home project!

That's about when its BOINC clients were reoptimized by hand by some industrious coder out there into it?

Intel CPP compiler using Intel IPP 4.1 libs made versions that KICKED THE HECK out of MSVC++ built ones in Fpop (floating point) data tests, which IS what SETI@Home works on mostly.

MSVC++ did produce better integer code though (means faster loops mostly imo (for most apps, unless a number smasher like Excel etc.) because loops aren't 2.5 - x.5 "iterations, lol, but solid integers like 2-n loops - most apps use integer data more than anything though usually, & the loop example's decent enough)... apk