Windows, Hyper Threading, and 3DMark 2006.

kid41212003 · May 11, 2009

Here are some 3DMark 2006 results with Hyper Threading on and off on different OSes.

--------------Windows XP-Windows Vista-Windows 7 (All x64)
[4GHz HT off]
SM 2.0 -------|8789|------|8939|--------|8950|
SM 3.0 -------|9328|------|9468|--------|9460|
CPU ----------|7030|------|6113|--------|6238|
[4GHz HT on]
SM 2.0 -------|8764|------|8492|--------|8959|
SM 3.0 -------|8857|------|9446|--------|9469|
CPU ----------|7549|------|6725|--------|6628|
[4.2GHz HT off]
SM 2.0 -------|8883|------|9054|--------|9056|
SM 3.0 -------|9400|------|9515|--------|9516|
CPU ----------|7373|------|6378|--------|6465|

Summary:
Windows XP has the best CPU score, SM 3.0 point decrease with HT on (9328->8857)
Windows Vista SM 2.0 point decrease with HT on (8939->8492)
There is no different on Windows 7 except CPU score, and it has the best SM 2.0 and 3.0 points.

XP has the best overall score because the CPU score is alot higher than others.
There is lil different on Windows 7 when OC the CPU to 4.2GHz, this could be GPU limited not CPU. If this is true, with a faster GPU, it's quite possible that with HT on and Windows 7, the score will come out on top (proving HT on + Windows 7 = best performance).

Screenshots:

Windows XP

Windows Vista

Windows 7

Thrackan · May 11, 2009

Nice comparison, thanks!

lemonadesoda · May 11, 2009

This is potentially very interesting analysis. However, please post the full 3Dmark06 score, not just the breakdown.

Reason: under the new GDI model of Vista and W7, there is a change in the balance of that the CPU is doing. For all we know, the CPU scores shown my 3Dmark06 my be misrepresentative of what is actually going on, since the CPU score is based on some kind of "subtraction algorithm".

The only true test is to run a game with a fixed timedemo and see the average FPS. Unless we do that, we cannot assume the 3Dmark06 scores really mean "games run faster under XP"

Asylum · May 11, 2009

Yes that would be my belief also...You would have to run something like Crysis benchmark tool on the different OS like 3 times each and get the average to have a acurate sum!!

h3llb3nd4 · May 11, 2009

Ok, XP for Benching and Win 7 for normal tasks

kid41212003 · May 11, 2009

Full score? Isn't it showing on those screenshots?

Graphic tests are represent games, not all games were built the same way like 3dmark 2006, but at least, it's should be the same.
The reason games usually act differently than 3dmark tests are because they were built for a different reason.
One for benchmark, and another for gaming. Of course, programmers always want their games light and run fast, but with benchmarks, it's different, it takes everything it can to gain more FPS.

I will do Company of Heroes benchmark, since it has a built-in benchmark.

kid41212003 · May 11, 2009

And that's why I put up all the SM scores, instead of Overall score, duh!

lemonadesoda · May 11, 2009

Perhaps my earlier comment wasnt clear.

If you want to consolidate the data (in the table you posted up top), to help people review analyse your empirical results, then consolidate it all. Don't skip the "final score".

Your data is like an olympic scoreboard with laptimes for the first 3 laps, but then omitting the final result!

Chicken Patty · May 11, 2009

thanks kid, this is a really good test!! Thanks for sharing and taking your time to do this.

kid41212003 · May 12, 2009

lemonadesoda said:
Perhaps my earlier comment wasnt clear.

If you want to consolidate the data (in the table you posted up top), to help people review analyse your empirical results, then consolidate it all. Don't skip the "final score".

Your data is like an olympic scoreboard with laptimes for the first 3 laps, but then omitting the final result!

What i'm trying to do here is not the best 3dmark 2006 overall score, but how Hyper Threading affect the graphic tests (SM), which is more importance, because it's about the frame per second. Better cpu score doesn't mean better FPS.

Do you know why you always get better CPU score on Windows XP? It's not because it runs faster, but it's because when the CPU test is running, everything else (backgrounds programs) is frozen.
Do you know why when you installing drivers, running prime95 on XP, there's always freezing, choppy, hiccups?
And do you know why it's not happen with Windows Vista and Seven? It's pretty much about how Windows share cpu resources between programs. And I think Vista and 7 are smarter about that, zero hiccups or freezing when multi-tasking.

About Company of Heroes.
There is something wrong with the new WHQL driver for XP, the game could not run because it thinks the card doesn't meet minimum requirement :twitch:

.

lemonadesoda · May 12, 2009

Try something like lightsmark 2008 just for starters

System Name	Acer Aspire V3-771G-53218G75Maii
Processor	Core i5 3210M (2,5-3,1Ghz)
Memory	8GB DDR3 SODIMM
Video Card(s)	Geforce GT650M
Storage	Samsung 830 256GB - 750GB Toshiba drive
Software	Windows 7 x64 Home Premium (non-acer-bloatware)

System Name	ICE-QUAD // ICE-CRUNCH
Processor	Q6600 // 2x Xeon 5472
Memory	2GB DDR // 8GB FB-DIMM
Video Card(s)	HD3850-AGP // FireGL 3400
Display(s)	2 x Samsung 204Ts = 3200x1200
Audio Device(s)	Audigy 2
Software	Windows Server 2003 R2 as a Workstation now migrated to W10 with regrets.

System Name	KILLER
Processor	Intel Core i7 4790K
Motherboard	MSI Z87 G45 Gaming
Cooling	Water Cooling
Memory	2x8GB Gskill 2400mhz
Video Card(s)	EVGA GTX 980
Storage	SSD Samsung 500GB
Display(s)	24" Asus 144hz
Case	Cooler Master HAF
Audio Device(s)	X-Fi Platinum
Power Supply	Antec 1200 Watt Modular
Software	Windows 10 64 Bit
Benchmark Scores	3DMark Vantage 51342

System Name	My mobo is Laughing at me
Processor	E7400
Motherboard	P5KPL-E Bios flashed to 0601 (Piece of poo!!)
Cooling	Thermalright Ultra 120a
Memory	GENERIC 2 GB DDR2 800
Video Card(s)	NONE!! Beat that!!
Storage	500GB SAMSUNG SATAII, 250GB SAMSUNG SATAII and o'l crappy 4gb maxtor
Display(s)	ACER X223W Q
Case	AEROCOOL ZERODEGREE (planning to mod)
Audio Device(s)	REALTEK ONBOARD
Power Supply	GIGABYTE 460W
Software	Win 7 x86 build 7022
Benchmark Scores	Super Pi 1m 17.000 :(

System Name	ICE-QUAD // ICE-CRUNCH
Processor	Q6600 // 2x Xeon 5472
Memory	2GB DDR // 8GB FB-DIMM
Video Card(s)	HD3850-AGP // FireGL 3400
Display(s)	2 x Samsung 204Ts = 3200x1200
Audio Device(s)	Audigy 2
Software	Windows Server 2003 R2 as a Workstation now migrated to W10 with regrets.

Processor	AMD Ryzen 7 7800X3D
Motherboard	ASUS ROG Crosshair X670E Hero
Cooling	ARCTIC Liquid Freezer III 360 A-RGB // X12 Phanteks D30-120 D-RGB Fans
Memory	G.SKILL Trident Z5 Neo RGB Series (AMD Expo) DDR5 RAM 32GB (2x16GB) 6000MT/s CL30
Video Card(s)	ZOTAC Gaming GeForce RTX 4090 AMP Extreme AIRO
Storage	Samsung Pro 980 2TB NVMe (OS and Games) // WD Black 10TB HDD (Storage)
Display(s)	SAMSUNG 34-Inch Ultrawide 75Hz (Desk Games) // SAMSUNG 49-Inch Ultrawide 144Hz (Sim Racing Rig)
Case	Phanteks NV7
Power Supply	Seasonic Vertex GX-1200, 1200W 80+ Gold, ATX 3.0
Software	Windows 11 Pro

Windows, Hyper Threading, and 3DMark 2006.

New Member