Here's how the GTX 560 scales with different CPUs:

Pentium D 830 (3GHz)

Athlon X2-3800 (2GHz)

Pentium E2220 (2.4GHz)

Opteron 185 (OC 3GHz)

Core 2 Duo E6600 (2.4GHz)

Core 2 Quad Q6600 (OC 3GHz)

Let's take a look at how much Intel IGPs have improved over the years.

Celeron 1.7GHz, Intel Extreme (i845GL), 1GB DDR266

Atom N470 (1.83GHz), GMA3150, 2GB DDR2-667

C2D T5600 (1.83GHz), GMA950, 4GB dual-channel DDR2-667

C2D E7500 (2.93GHz), GMA X4500HD, 4GB dual-channel DDR3-1066

Atom X7 Z8700 (Surface 3), Intel HD iGPU, 4GB dual-channel DDR3-1600:

i5-3317U (Surface Pro first gen), Intel HD 4000, 4GB dual-channel DDR3-1600:

Here's a P3-1000 @ 1125 with an x800XT. Not sure why the CPU score is so low with this one. Even the PIII-866 w/ FX5200 gets a higher CPU score.

Link to the 3DMark 2001 page (just for reference) and added this page to the Interesting Vogons Threads page on Vogons wiki (even though this thread is kinda short 😜)

Dunno if there was already a thread about 3Dmark 2003, but the fact that it's more a GPU benchmark is interesting in itself and chances are this thread will be picking up in the future.

It's not my result, found it here http://www.overclockers.at/benchmarks/k6-3_un … ighlight=radeon

Definitely a GPU benchmark.

AMD K6-3, 400MHz (4.0 x 100 MHz)
Tyan Trinity 100AT (Via MVP3 Chipset)
2x128MB Kingston PC100
Radeon 9800Pro Catalyst 4.2
D3D & OGL Quality: Balanced
OS: W2k Sp3
IBM DTLA-307030

3DMark 2003 score: 2754

Pentium D-965 Extreme @ 4.66GHz, Quadro FX4800, 4GB dual-channel DDR2-8500 CL5: 39021

dottoss wrote:

Pentium D-965 Extreme @ 4.66GHz, Quadro FX4800, 4GB dual-channel DDR2-8500 CL5: 39021

As a fellow owner/user of the legendary heater know as Pentium Extreme Edition 965 (OC) I HOPE you are using water cooling and have a badass motherboard to run it @1,55V 😁
Nice score, but how much of that ~1,55V stays after putting load on CPU ?

For comparison, here are two of my own scores :
"Cheap" MB (w/Win 7 SP1 x64) :

And going mad (w/PCI-e 2.0/DDR3/Win 10 x64) :

PS.

Another nice "feature" of 3DMark03 is that it doesn't take a performance hit with newer versions of Windows.

That's really not true for Win 10 x64... but I guess it's still "in beta" and I'm just a "beta tester" 😜

s a fellow owner/user of the legendary heater know as Pentium Extreme Edition 965 (OC) I HOPE you are using water cooling and have a badass motherboard to run it @1,55V 😁
Nice score, but how much of that ~1,55V stays after putting load on CPU ?

A fair amount of v-drop: however, i have disabled the load Load-Line Calibration.

I can OC it up to 5ghz but then it requires around 1.7v and it will eventually shutdown thermally. It`s stable up to around 4.9GHz at somewhat safe voltages but stable 5.0GHz seems to be out of reach for this particular 965.

Cooling: kelvin s36

Motherboard: Asus P5Q-Deluxe

And going mad (w/PCI-e 2.0/DDR3/Win 10 x64)

I`ve been thinking of moving to either x38 or x48, what do you recommend? The P45 chipset is fine, but the X is just that better.

Well, "X" series simply has two full PCI-e x16 slots available for SLI/CFX.
P35 doesn't have PCI-e 2.0 and X48 is a higher binned X38.
Best MB for this chip is the one with best VRM cooling - that's why I bought Rampage Extreme 😀

1,7V for 5GHz ?
Holy...
I simply put in BIOS 1,66V on Vcore (Vdroop ~0,1V), 1,35V on VTT and 1,65V on PLL 😀
With 1,55V on NB and 1,65V on RAM, I can do that Win 10 score.
And like You I have to turn off my Load Line Calibration (it gave me black screens...).

Have you tried going 1600MHz+ and lowering multiplier ?
Also : ALWAYS check VRM temps while in benchmark.
My P5K64-WS is a solid P35 board, but what it lacks is VRM cooling (at least when paired with this monster after OC).

Only thing that "X" has are two electricaly "x16" PCI-e 2.0 slots.

PS. I cool mine with Thermalright 120 Ultra Extreme with 4k Delta FFB1212EHE Fan on top 😉
My max core/fsb clock for now is 5,1GHz at 425MHz [1700MHz] : LINK

falloutboy wrote:

It's not my result, found it here http://www.overclockers.at/benchmarks/k6-3_un … ighlight=radeon […]
Show full quote

It's not my result, found it here http://www.overclockers.at/benchmarks/k6-3_un … ighlight=radeon

Definitely a GPU benchmark.

AMD K6-3, 400MHz (4.0 x 100 MHz)
Tyan Trinity 100AT (Via MVP3 Chipset)
2x128MB Kingston PC100
Radeon 9800Pro Catalyst 4.2
D3D & OGL Quality: Balanced
OS: W2k Sp3
IBM DTLA-307030

3DMark 2003 score: 2754

That's awesome! I just wish he would have tried the CPU tests on that.

agent_x007 wrote:

For comparison, here are two of my own scores :
"Cheap" MB (w/Win 7 SP1 x64)

66.6K--that's a great score for Netburst! My 3GHz Opteron 185 could only do 52.5K with a GTX 560 and WinXP. Now I'd like to retest it with a GTX 680 and Win7. For some reason XP always seems to produce lower CPU scores (but not as bad as Win10! 😳 )

Well, "X" series simply has two full PCI-e x16 slots available for SLI/CFX.
P35 doesn't have PCI-e 2.0 and X48 is a higher binned X38.
Best MB for this chip is the one with best VRM cooling - that's why I bought Rampage Extreme

Thanks, I will look for an Asus rampage. The CPU draws a hefty amount of current during full load, so I guess the more VRM cooling it has, the more it can push before it being deteriorating to the CPU and/or the MB. The P5Q-Deluxe has no setting for the Command rate and defaults to 2T, does the Rampage has this in BIOS?

agent_x007 wrote:

Well, "X" series simply has two full PCI-e x16 slots available for SLI/CFX. P35 doesn't have PCI-e 2.0 and X48 is a higher binne […]
Show full quote

Well, "X" series simply has two full PCI-e x16 slots available for SLI/CFX.
P35 doesn't have PCI-e 2.0 and X48 is a higher binned X38.
Best MB for this chip is the one with best VRM cooling - that's why I bought Rampage Extreme 😀

Have you tried going 1600MHz+ and lowering multiplier ?
Also : ALWAYS check VRM temps while in benchmark.
My P5K64-WS is a solid P35 board, but what it lacks is VRM cooling (at least when paired with this monster after OC).

Yes, I've tested it, it borks out as soon as I hit 416 FSB, and from 410 FSB to 416 FSB I need to increase the volt like crazy. I think it's the CPU that is the limit, but let's see when I get ahold of an Rampage board 😀

dottoss wrote:

Thanks, I will look for an Asus rampage. The CPU draws a hefty amount of current during full load, so I guess the more VRM cooling it has, the more it can push before it being deteriorating to the CPU and/or the MB. The P5Q-Deluxe has no setting for the Command rate and defaults to 2T, does the Rampage has this in BIOS?

VRM cooling can be misleading tho.
For eample, my P5K64-WS has a pretty good VRM cooling (on pics).
BUT, because top one is mounted on plastik pins (instead of metal screws) and has no heatpipes, it cannot make good contact with MOSFET's underneath when I put heavy cooling on my CPU (PCB gets bend over the CPU socket area to the point of visible gap can be seen between VRM heatsink and MOSFET's themselves).
That's how I had throttling on CPU @ 60C on it.

More current = more heat - true.
But, more MOSFET's = less heat (because current in split between all of them).
MOSFET temps are critical when they get around 90-100C.
If they aren't properly cooled, they can blow up under load.
That's why I always check temp of PCB edge closeset to MOSFET bank.
If I can't hold a finger to it for more than a 10 sec. MOSFET's are probably overheating.
^I had this problem with P5K64-WS and Pentium XE 965@5GHz.
Since CPU couldn't drop speed or Vcore, it used "STOP" instruction to not overheat it.
Performance dropped, but VRM didn't blew up 😀

@up Did you play around with Strap's and Performance Level and FSB Voltage (VTT) ?

Standard Def Steve wrote:

agent_x007 wrote:

For comparison, here are two of my own scores :
"Cheap" MB (w/Win 7 SP1 x64)

66.6K--that's a great score for Netburst! My 3GHz Opteron 185 could only do 52.5K with a GTX 560 and WinXP. Now I'd like to retest it with a GTX 680 and Win7. For some reason XP always seems to produce lower CPU scores (but not as bad as Win10! 😳 )

Hmm... I don't know ~66k is a bit low.
Give me a sec...
...
OK, I think I got this 😀

74k seems about right, won't you agree ?

agent_x007 wrote:

74k seems about right, won't you agree ?

That's actually fairly amazeballs for a PEE (lol...pee). Both the score and the 5.1GHz.

I paired up my GTX970 with a Phenom X6 1090T at 4069MHz. Score was a bit lower than I expected. I didn't think a Ph2 @ 4GHz would bottleneck a 970 that much in 3D'03.

Dredged up from the massive pit of benchmark results on my file server, here's a crappy Geforce GT635 with a Core i7 3770 (3.4GHz/3.9Turbo).
The CPU score's a little on the low side (even less than the 4GHz Phenom II I posted above), but that's probably because I was running it on the Win10 Technical Preview.

My result with a Core2 Duo 3ghz 2GB DDR2 Geforce 7950GT.

Going to try 3DMark06 next, may not be pretty.

Messing around with a slow 4CoreDual-SATA 2 board.
C2D E6600 2.4GHz/1066, 4CoreDual-SATA2 (VIA PT880 Pro) 2GB DDR2-533 3-3-3-9, onboard audio, XP SP3

With GeForce 6800GT AGP:

With Radeon x1950Pro AGP:

With Geforce 9800GT PCIe (1.1 @ x4)