Post Your Framerate Benchmark Here

[m4xc4v413r4]

A benchmark in WoW is impossible to be comparable since you can't really do the same exact thing twice.... so what's the point in this?

[vesseblah]

Actually you can do the same thing twice. Just can't measure raid performance with it.
http://www.wowwiki.com/Taxi_Time_Test

[Cilraaz]

A benchmark in WoW is impossible to be comparable since you can't really do the same exact thing twice.... so what's the point in this?

Yes, you are right that no two benchmarks on the same system will yield the same framerate result every time. Then again, I don't get the same numbers from my Counter-Strike: Source tests every time. 3DMark06 doesn't give me the same value every time. PCMark Vantage and 3DMark Vantage don't give me the same values every time. Benchmarks are not perfect in any scenario. You can only do a best-case test, whether it be controlled (/timetest over a certain flight) or real-world (multiple raid runs). Multiple tests will help even out the randomness and imperfection, but even a single test is close enough, especially when discounting blatant outliers.

This thread isn't intended to show precise performance numbers, but rather what a person should roughly expect from a specific setup.

[None]

I forgot to mention I also have my northbridge overclocked to 2.6ghz, that might skew the performance difference from the stock Phenom II x3 720 and my Phenom II x4 955. Also during the second test at 4.0ghz Fraps seemed to lock at 250fps, the counter just seemed to pause at exactly 250 and the World of Warcraft counter kept jumping; either way, the Fraps log doesn't match the Timetest log.

Any more of these?

I'm going to try for 3ghz northbridge tonight, hopefully I can get it stable enough for a benchmark.

[llDemonll]

System Specs including:
CPU: Intel Core i7 920 (2.66ghz)
GPU: Sapphire Radeon HD 4870 256-bit 512MB
RAM: G.Skill 3GB (3x1GB) DDR3 1333 PC10666 7-7-7-24
MOBO: Gigabyte GA-EX58-UD3R X58 ATX Motherboard
HD: Western Digital Caviar Black 7200rpm 500GB

WoW Settings including:
Resolution: 1440x900 Full-Screen Windowed
Multisampling Rate (AA): 2x

Graphic Settings
V-Sync Setting: Enabled
Texture Filtering: 4x Antistropic
Triple Buffering Setting (if V-Sync is enabled): Disabled
Textures: ALL Highest Setting
Environment: ALL Highest Setting
Effects: ALL Highest Setting EXCEPT Shadows is on High not Ultra
Zone or Instance: 25M ICC Heroic Dreamwalker, Gauntlet, Trash (both packs at once) and Sindragosa

Benchmark Results including:
Frames Captured: 71028 (23:45 Minutes)
Minimum Framerate: 9 FPS
Maximum Framerate: 62 FPS
Average Framerate: 50 FPS


The readable FPS log file shows 14 frames (captured each second in this case) that were under 30 FPS; [9, 10, 10, 10, 11, 11, 15, 18, 19, 19, 24, 26, 27]

[Deleted]

Nice Demon. But i thought the the HD 4870 is good enough to have AA on more than 2x? (more of a question than a statement)

[Asera]

It probably would have trouble at 4x if he ran at 1080p due to that 4870 being only 512mb (I didn't even know they made 512mb versions?)
At 1440x900 though it shouldn't.. :s

[llDemonll]

ya, and I like my pixel borders. it's been at 2x for..couple years now and i just have never tried higher. there isnt a ton of difference between the two, and i'm not going to notice any difference while raiding anyways so i just stick with 2 . and ya, the 19" monitor helps the fps a bit :P

[Deleted]

Flight on Taxi - River's Heart (Sholazar Basin) to Dalaran

PC:
Cpu: AMD Phenom II x6 at 2.8ghz, 3.2ghz turbo
Ram: 6GB Kingston DDR3 HyperX at 2000mhz
GPU: XFX Radeon 5870 1gb ddr5 (875mhz gpu clock and 5.2ghz memory clock (stock values))
HDD: 7200rp 1Terabye drive, diffrent drive then Windows (which is on a identical disc) (Samsung spinpoint f3)

Running everything at highest and multisample on 8x and 1920x1080 resolution windowed mode

Results: (Min, Max, Avg)
39, 160, 67.077

[Deleted]

Will probably post some benchmarks of an Athlon II x3 445 when i get it.
Just so people can see what kind of framerates lower end processors get. (Since this topic is mostly full of Phenom II's and I5's)

[Wries]

When comparing DX 9 and DX 11 mode on my mobility radeon 5650 I ran into a problem with timetest and that was that it capped max framerate to 66.667 fps even though I was above 100 fps a couple of times during the flight. I assume average fps gets ruined as well when the maximum fps reported by the tool is bogus.

This happened on both DX 9 and DX 11, and only went away on the fourth try when I was running DirectX 11 and had the card on a hefty overclock. Vertical sync was of course turned off.

The minimum fps was weird as well, reporting 6-8 fps even though fps weren't close to that at any point of the timetest.

Have anyone else had this?

[Deleted]

Still downloading the patches for WoW. But after that ill Timetest this fresh system from edge to edge kalimdor and eastern kingdoms.
People must be curious how a lower end (compared to phenom II) Athlon II x4 @ 3.1ghz performs.

[khanggaroo]

I play with an 8800GTS 640MB at 1680x1050 on high perfectly fine. The new patch seems to have optimized how WoW runs or something.

[denominator]

Figured it's the end of the expansion why not take a last ode by flying across the continent, luckily I also ran a benchmark while doing so. Which should help at least someone I hope.

Intel core i7 920 (bloomfield at 4.2ghz on a 21x multiplier and 1.3v [core])
Asus EAH4890 (Ati radeon 4890) at 975mhz
6 (3x2) Gigs OCZ gold at 584mhz (CAS 8.0)
Resolution: 1920x1080
Multisampling Rate: 8x
V-Sync disabled (for benchmarking only)
Triple Buffering disabled for the test because v-sync was off.
Windowed but maximized
Full Screen Glow, Specular Lighting, Projected Textures enabled. max everything (except shadow - high).
For the test I took a flight path from Warsong Hold to Vengeance Landing

[None]

I've been rerunning the Tomshardware test (Crushblow to The Krazzworks) and have been getting completely different numbers. On average I've been getting 78.301 frames per second using identical settings to TomsHardware (I think. They might have used 1xMSAA with 16xAF, I used 8xMSAA and 4x AF), and with my Phenom II x4 955 at STOCK (Just for reference to Tomshardware, when I have overclocked results I'll post my full results.)

I don't know what the hell they did to make a Phenom II x6 get lower framerates at a higher frequency. Thuban is identical to Deneb cores, using a 6 core Thuban CPU should give identical results to a Deneb CPU at the same frequency.

This test was with a Phenom II x4 955 at stock (3.2ghz with 2ghz Northbridge, 1600mhz 8-8-8-24 2T) I also borrowed a 1680x1050 monitor for testing just to compare directly to their CPU tests.
I'll do a video next time as well just to show that there's literally no slow down or choppiness, the low minimum frame was a bit surprising because the framerate counter never dropped that low, nor did it ever show the maximum framerate that the timetest reported.
I'll also be doing it without timetest and will use Fraps to monitor the framerate so I can have NPC and players present in the test. I may also pull my 5770 out of my other computer to test the difference between a 470 and that. In the next few tests I'll also do various levels of overclocking, including RAM, northbridge and CPU frequency. Unfortunately time of day is now a factor in benchmarking World of Warcraft since Sunshafts are only available in certain zones and certain times of day.

Instead of rambling on here's what the first test showed:
AMD Phenom II x4 955 @ 3.2ghz/2ghz Northbridge
G.Skill Ripjaws 8-8-8-24 2T @ 1600mhz

I'll be adding more later when the servers come back up.

Also note that TomsHardware did their tests on the Beta. I am in no way claiming their results to be biased or invalid.

[Cilraaz]

I don't know what the hell they did to make a Phenom II x6 get lower framerates at a higher frequency. Thuban is identical to Deneb cores, using a 6 core Thuban CPU should give identical results to a Deneb CPU at the same frequency.

Except that there hasn't been a benchmark yet that didn't show a Phenom II x4 ahead of a Phenom II x6. Regardless of how the performance should be in theory, the x4 has repeatedly outperformed the x6 in these benchmarks.

[None]

Except that there hasn't been a benchmark yet that didn't show a Phenom II x4 ahead of a Phenom II x6. Regardless of how the performance should be in theory, the x4 has repeatedly outperformed the x6 in these benchmarks.

My guess would be AMDs turbo causing issues, it's likely dropping down to 3 cores at 3.4ghz or something, in which case any Phenom II x4 at 3.0ghz or above should beat it. If Turbo were disabled and both were clocked the same in WoW it would definitely be identical, there would be no reason why it shouldn't be.

Edit: I have most of the information now, I just need to sort through it and graph it.
Doing tests at 3.4ghz, 3.8ghz, and 4.0ghz for the most common clocks on a Phenom II x4. 3.4ghz is just for reference when buying a Phenom II x4 965 at stock, and I was too lazy to do 3.5ghz for the 970. I'll also include stock northbridge up to 2.6ghz in increments.

[ikshaar]

Computer
Intel(R) Core(TM) i5 CPU 760 @ 2.80GHz
ASUSTeK Computer INC. P7P55D-E PRO
2x4GB DDR3 Corsair XMS 1600
MSI N460GTX Cyclone 1GB
1920x1200

nVidia
AAx4 override
Supersample
Aniso. in game

WoW
Fullscreen
Full glow ON (/console ffxGlow 1)
Ultra setting except:
Shadow Good
Anisotropic Filtering 4x
Multisampling 4x

Flying SW to Andorhal with /timetest command
so DX11: average 106 and DX9 (tested twice - hence two lines) average 102-104.



FPS obtained with FRAPS. timetest reported false results twice... beware.

[TanAxys]

I built a new rig, so I thought I'd give this a try

The bottleneck, so to speak, for these tests is my monitor. It maxes out at 1440x900.

System Specs:
AMD Phenom II x3 720 @ 2.8 (200x14) stock
MSI 5670 1GB (775MHz, 1010MHz VRAM) stock http://www.newegg.com/Product/Produc...82E16814127489
G.Skill DDR3 4GB CL9 @ 1333MHz (PC10666) stock http://www.newegg.com/Product/Produc...82E16820231274

Settings:

Maxed everything, except shadows on good

Here's the standard Time Test (Crush to Kraz)


I also did another, that basically anyone could do. Ratchet to Gadgetzan. Has a lot of water to push the card a bit.


The big thing I've noticed with this set up is a split second stutter when mounting the taxi. I think that's where the frame rate dips into single digits. That's a non factor for my game play tho. I've had my cpu unlocked before, so I think I'll run the test again a few times with 4 cores, and then maybe again with some OC thrown in, to see what kind of a difference the cpu makes in these tests.

[Cilraaz]

I tend to not use /timetest. My maximum framerates have hit 550+ according to Blizzard's output, when fraps says I topped out at 150-200. It also shows a lower low than fraps. The average is usually pretty close, though.