Monday, March 31, 2008

Windows Vista 64

I have just loaded Vista 64 and it seems to be working well. I had the same problem with the network drivers where the files downloaded from Dlink's website caused the machine to hard lock.

To fix it this time I just plugged in one of the Ethernet ports and let windows update the driver from windows update. It worked great and just installed the updated Atheros driver.

I have been impressed with Vista 64's hardware support so far. I will begin installing programs and games later on and get back with some Crysis benchmarks.

Thursday, March 27, 2008

Crysis DX9 CPU Benchmark #2 on Vista32 w/174.44

Forceware 174.44 1920x1200 "high" no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 67.50s, Average FPS: 22.22
Min FPS: 6.38 at frame 478, Max FPS: 38.60 at frame 120
Average Tri/Sec: -3823367, Tri/Frame: -172059
Recorded/Played Tris ratio: 4.86
!TimeDemo Run 1 Finished.
Play Time: 65.94s, Average FPS: 22.75
Min FPS: 6.34 at frame 524, Max FPS: 38.89 at frame 114
Average Tri/Sec: -4704593, Tri/Frame: -206829
Recorded/Played Tris ratio: 4.04
!TimeDemo Run 2 Finished.
Play Time: 69.35s, Average FPS: 21.63
Min FPS: 3.63 at frame 471, Max FPS: 39.38 at frame 110
Average Tri/Sec: -4440437, Tri/Frame: -205299
Recorded/Played Tris ratio: 4.07
!TimeDemo Run 3 Finished.
Play Time: 66.21s, Average FPS: 22.65
Min FPS: 3.63 at frame 471, Max FPS: 39.42 at frame 114
Average Tri/Sec: -4648335, Tri/Frame: -205189
Recorded/Played Tris ratio: 4.07
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX10 CPU Benchmark #2 on Vista32 w/174.44

Forceware 174.44 1920x1200 "high" no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 87.41s, Average FPS: 17.16
Min FPS: 4.33 at frame 472, Max FPS: 26.33 at frame 108
Average Tri/Sec: -3142573, Tri/Frame: -183121
Recorded/Played Tris ratio: 4.56
!TimeDemo Run 1 Finished.
Play Time: 78.15s, Average FPS: 19.19
Min FPS: 4.33 at frame 472, Max FPS: 26.90 at frame 107
Average Tri/Sec: -3931207, Tri/Frame: -204828
Recorded/Played Tris ratio: 4.08
!TimeDemo Run 2 Finished.
Play Time: 76.97s, Average FPS: 19.49
Min FPS: 4.33 at frame 472, Max FPS: 26.90 at frame 107
Average Tri/Sec: -4025619, Tri/Frame: -206576
Recorded/Played Tris ratio: 4.04
!TimeDemo Run 3 Finished.
Play Time: 82.30s, Average FPS: 18.23
Min FPS: 3.28 at frame 472, Max FPS: 26.96 at frame 108
Average Tri/Sec: -3813943, Tri/Frame: -209252
Recorded/Played Tris ratio: 3.99
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX9 CPU Benchmark on Vista32 w/174.44

Forceware 174.44 1920x1200 "high" no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 35.81s, Average FPS: 41.88
Min FPS: 28.82 at frame 836, Max FPS: 57.89 at frame 341
Average Tri/Sec: -57299224, Tri/Frame: -1368077
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 1 Finished.
Play Time: 31.60s, Average FPS: 47.47
Min FPS: 28.82 at frame 836, Max FPS: 62.94 at frame 123
Average Tri/Sec: -64792032, Tri/Frame: -1365008
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 31.84s, Average FPS: 47.11
Min FPS: 28.33 at frame 1103, Max FPS: 62.94 at frame 123
Average Tri/Sec: -64218328, Tri/Frame: -1363039
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 31.73s, Average FPS: 47.28
Min FPS: 28.33 at frame 1103, Max FPS: 62.94 at frame 123
Average Tri/Sec: -64304580, Tri/Frame: -1360205
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX10 CPU Benchmark on Vista32 w/174.44

Forceware 174.44 1920x1200 "high" no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 52.24s, Average FPS: 28.71
Min FPS: 15.91 at frame 826, Max FPS: 37.71 at frame 307
Average Tri/Sec: -39302096, Tri/Frame: -1368883
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 1 Finished.
Play Time: 45.99s, Average FPS: 32.62
Min FPS: 15.91 at frame 826, Max FPS: 40.64 at frame 123
Average Tri/Sec: -44429668, Tri/Frame: -1362193
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 46.88s, Average FPS: 32.00
Min FPS: 15.91 at frame 826, Max FPS: 40.84 at frame 118
Average Tri/Sec: -43629132, Tri/Frame: -1363598
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 45.75s, Average FPS: 32.79
Min FPS: 15.91 at frame 826, Max FPS: 40.84 at frame 118
Average Tri/Sec: -44760128, Tri/Frame: -1365090
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX9 GPU Benchmark on Vista 32bit w/174.44

1920x1200 on "high" no AA

This is pretty good for Crysis. There must be a lot of room left in the software for optimization. Every time a new driver comes out there is a 5-15% increase in Crysis performance.

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 47.76s, Average FPS: 41.88
Min FPS: 20.82 at frame 147, Max FPS: 54.40 at frame 858
Average Tri/Sec: -27000780, Tri/Frame: -644770
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 42.48s, Average FPS: 47.08
Min FPS: 20.82 at frame 147, Max FPS: 56.69 at frame 95
Average Tri/Sec: -30233604, Tri/Frame: -642203
Recorded/Played Tris ratio: -1.43
!TimeDemo Run 2 Finished.
Play Time: 42.41s, Average FPS: 47.16
Min FPS: 20.82 at frame 147, Max FPS: 56.69 at frame 95
Average Tri/Sec: -30265908, Tri/Frame: -641785
Recorded/Played Tris ratio: -1.43
!TimeDemo Run 3 Finished.
Play Time: 42.45s, Average FPS: 47.11
Min FPS: 20.82 at frame 147, Max FPS: 58.45 at frame 1013
Average Tri/Sec: -30382418, Tri/Frame: -644887
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX10 GPU Benchmark on Vista 32bit w/174.44

Here is the first benchmark with Forceware 174.44. It looks like it makes a difference.

1920x1200 "High" no AA

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 69.74s, Average FPS: 28.68
Min FPS: 17.85 at frame 142, Max FPS: 34.94 at frame 992
Average Tri/Sec: -18478744, Tri/Frame: -644373
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 65.19s, Average FPS: 30.68
Min FPS: 17.85 at frame 142, Max FPS: 39.54 at frame 107
Average Tri/Sec: -19702422, Tri/Frame: -642195
Recorded/Played Tris ratio: -1.43
!TimeDemo Run 2 Finished.
Play Time: 62.87s, Average FPS: 31.81
Min FPS: 17.85 at frame 142, Max FPS: 39.79 at frame 61
Average Tri/Sec: -20433948, Tri/Frame: -642297
Recorded/Played Tris ratio: -1.43
!TimeDemo Run 3 Finished.
Play Time: 62.89s, Average FPS: 31.80
Min FPS: 17.85 at frame 142, Max FPS: 39.90 at frame 64
Average Tri/Sec: -20485420, Tri/Frame: -644191
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

New Forceware v174.74 released today.

All the vista benchmarks so far have been using Forceware v169.44 will the new driver version make any difference. When my download is complete I'll let you know.

3Dmark06 score of 16012


Here is my new 3Dmark06 score on Vista32. Its a little bit lower than the XP score.

DX9 vs DX10 on Vista32

It seems that there is a 15-40% performance hit for using DX10 on Crysis. Thats a lot! I guess I will be gaming with DX9 for Crysis until I get some more powerful hardare or until DX10 becomes more optimized.

I think I am going to try these same tests on Vista64 and see if there is any difference.

Crysis DX9 CPU Benchmark #2 on Vista32

1920x1200 on "high" no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 78.42s, Average FPS: 19.13
Min FPS: 7.23 at frame 477, Max FPS: 30.71 at frame 121
Average Tri/Sec: -3581995, Tri/Frame: -187267
Recorded/Played Tris ratio: 4.46
!TimeDemo Run 1 Finished.
Play Time: 72.31s, Average FPS: 20.74
Min FPS: 6.70 at frame 478, Max FPS: 31.61 at frame 91
Average Tri/Sec: -4334965, Tri/Frame: -208966
Recorded/Played Tris ratio: 4.00
!TimeDemo Run 2 Finished.
Play Time: 72.66s, Average FPS: 20.64
Min FPS: 4.74 at frame 469, Max FPS: 31.61 at frame 91
Average Tri/Sec: -4202438, Tri/Frame: -203564
Recorded/Played Tris ratio: 4.10
!TimeDemo Run 3 Finished.
Play Time: 71.48s, Average FPS: 20.98
Min FPS: 4.74 at frame 469, Max FPS: 31.61 at frame 91
Average Tri/Sec: -4304010, Tri/Frame: -205102
Recorded/Played Tris ratio: 4.07
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX10 CPU Benchmark #2 on Vista32

1920x1200 on "high" with no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 98.62s, Average FPS: 15.21
Min FPS: 5.52 at frame 476, Max FPS: 22.21 at frame 119
Average Tri/Sec: -2788838, Tri/Frame: -183354
Recorded/Played Tris ratio: 4.56
!TimeDemo Run 1 Finished.
Play Time: 87.05s, Average FPS: 17.23
Min FPS: 5.52 at frame 476, Max FPS: 22.21 at frame 119
Average Tri/Sec: -3505312, Tri/Frame: -203421
Recorded/Played Tris ratio: 4.11
!TimeDemo Run 2 Finished.
Play Time: 87.63s, Average FPS: 17.12
Min FPS: 5.52 at frame 476, Max FPS: 22.21 at frame 119
Average Tri/Sec: -3546740, Tri/Frame: -207196
Recorded/Played Tris ratio: 4.03
!TimeDemo Run 3 Finished.
Play Time: 89.87s, Average FPS: 16.69
Min FPS: 5.52 at frame 476, Max FPS: 22.21 at frame 119
Average Tri/Sec: -3469375, Tri/Frame: -207861
Recorded/Played Tris ratio: 4.02
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

CPU Benchmakr DX9 vs DX10

It seems like there is a pretty big difference between DX9 and DX10 on this benchmark. It's not exactly apples to apples because DX10 looks better than DX9 does. It's almost like turning up the quality settings from "high" to "high and a half".

Crysis DX9 CPU Benchmark on Vista32

1920x1200 on "high" with no AA

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 42.86s, Average FPS: 35.00
Min FPS: 21.64 at frame 833, Max FPS: 48.86 at frame 339
Average Tri/Sec: -47882640, Tri/Frame: -1368232
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 1 Finished.
Play Time: 37.53s, Average FPS: 39.97
Min FPS: 21.64 at frame 833, Max FPS: 52.45 at frame 103
Average Tri/Sec: -54522460, Tri/Frame: -1364160
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 37.36s, Average FPS: 40.15
Min FPS: 21.64 at frame 833, Max FPS: 52.57 at frame 115
Average Tri/Sec: -54809312, Tri/Frame: -1365063
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 37.22s, Average FPS: 40.30
Min FPS: 21.64 at frame 833, Max FPS: 52.98 at frame 104
Average Tri/Sec: -55011328, Tri/Frame: -1365028
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX10 CPU Benchmark on Vista32

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 63.18s, Average FPS: 23.74
Min FPS: 14.65 at frame 1480, Max FPS: 30.88 at frame 79
Average Tri/Sec: -32500650, Tri/Frame: -1368918
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 1 Finished.
Play Time: 56.71s, Average FPS: 26.45
Min FPS: 14.65 at frame 1480, Max FPS: 34.93 at frame 104
Average Tri/Sec: -35948788, Tri/Frame: -1359178
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 56.15s, Average FPS: 26.72
Min FPS: 14.65 at frame 1480, Max FPS: 34.93 at frame 104
Average Tri/Sec: -36365512, Tri/Frame: -1361167
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 55.98s, Average FPS: 26.80
Min FPS: 14.65 at frame 1480, Max FPS: 34.93 at frame 104
Average Tri/Sec: -36461632, Tri/Frame: -1360657
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

per the usual all tests were run at 1920x1200 res with all settings on "high"

Wednesday, March 26, 2008

Crysis DX10 GPU Benchmark on Vista 32bit

Here are the same settings: 1920x1200 on "high" no AA, v1.2 with Forceware v169.44. As you can see there is a significant performance hit for using DX10.

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 85.39s, Average FPS: 23.42
Min FPS: 12.43 at frame 141, Max FPS: 30.01 at frame 980
Average Tri/Sec: -15165479, Tri/Frame: -647480
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 76.67s, Average FPS: 26.09
Min FPS: 12.43 at frame 141, Max FPS: 33.52 at frame 75
Average Tri/Sec: -16843580, Tri/Frame: -645710
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 2 Finished.
Play Time: 78.52s, Average FPS: 25.47
Min FPS: 12.43 at frame 141, Max FPS: 35.65 at frame 65
Average Tri/Sec: -16428409, Tri/Frame: -645011
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 3 Finished.
Play Time: 86.96s, Average FPS: 23.00
Min FPS: 12.43 at frame 141, Max FPS: 35.65 at frame 65
Average Tri/Sec: -14845135, Tri/Frame: -645439
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis DX9 GPU Benchmark on Vista 32bit

Here is the first benchmark from my Vista32 installation. Like I thought its a little slower than the XP benchmark.

This is the Crysis GPU benchmark on 1920x1200 with all settings set to "high".

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 58.60s, Average FPS: 34.13
Min FPS: 19.16 at frame 159, Max FPS: 47.08 at frame 899
Average Tri/Sec: -22016680, Tri/Frame: -645137
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 51.42s, Average FPS: 38.90
Min FPS: 19.16 at frame 159, Max FPS: 50.44 at frame 997
Average Tri/Sec: -25055082, Tri/Frame: -644107
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 2 Finished.
Play Time: 51.23s, Average FPS: 39.04
Min FPS: 19.16 at frame 159, Max FPS: 52.77 at frame 86
Average Tri/Sec: -25133784, Tri/Frame: -643853
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 3 Finished.
Play Time: 51.08s, Average FPS: 39.15
Min FPS: 19.16 at frame 159, Max FPS: 52.77 at frame 86
Average Tri/Sec: -25198608, Tri/Frame: -643597
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Tuesday, March 25, 2008

Vista32 is up and running!

I just got my copy of Windows Vista Ultimate SP1, and got it set up on the Quad. I had been refusing to buy Vista until SP1 came out, so when it did I went ahead and upgraded. Most of the installation went smoothly, but there was one hitch.

I have a DLink DWA-552 Extreme N wireless card set up on the Quad now and when I tried to install the v1.3 Vista drivers I got a hard lock up. I had been very clever or so I thought and had downloaded the Vista v1.2 drivers as well so I tried those. Same result. Hard lock.

A quick search of the forums on google showed a lot of other people having the same problem. Figures.

Ironically the fix for this problem is to use the v1.11 XP drivers. So that is what I am running on now. I hope that DLink will address this problem soon. It is crappy to buy a nice piece of hardware only to have the experience ruined by crappy drivers.

Oh well, it's working now. I'll get Crysis set up later this week and run some benchmarks so we can compare performance to XP. I don't expect to be pleased :( The price of progress.

Saturday, March 8, 2008

Crysis CPU Benchmark #2

Here is the CPU benchmark #2. It looks like this one remains CPU limited is not much different before or after the patch.

new (v1.2 and 169.44)
==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 70.49s, Average FPS: 21.28
Min FPS: 11.77 at frame 428, Max FPS: 32.93 at frame 102
Average Tri/Sec: -3722447, Tri/Frame: -174938
Recorded/Played Tris ratio: 4.78
!TimeDemo Run 1 Finished.
Play Time: 67.27s, Average FPS: 22.30
Min FPS: 6.63 at frame 473, Max FPS: 34.17 at frame 96
Average Tri/Sec: -4590217, Tri/Frame: -205867
Recorded/Played Tris ratio: 4.06
!TimeDemo Run 2 Finished.
Play Time: 66.59s, Average FPS: 22.53
Min FPS: 6.63 at frame 473, Max FPS: 34.17 at frame 96
Average Tri/Sec: -4616777, Tri/Frame: -204939
Recorded/Played Tris ratio: 4.08
!TimeDemo Run 3 Finished.
Play Time: 64.77s, Average FPS: 23.16
Min FPS: 6.63 at frame 473, Max FPS: 34.17 at frame 96
Average Tri/Sec: -4772555, Tri/Frame: -206084
Recorded/Played Tris ratio: 4.05
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

old (v1.1 and 169.28)
==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 76.64s, Average FPS: 19.57
Min FPS: 5.50 at frame 478, Max FPS: 32.37 at frame 102
Average Tri/Sec: -3510228, Tri/Frame: -179346
Recorded/Played Tris ratio: 4.66
!TimeDemo Run 1 Finished.
Play Time: 64.52s, Average FPS: 23.25
Min FPS: 5.50 at frame 478, Max FPS: 32.71 at frame 91
Average Tri/Sec: -4758319, Tri/Frame: -204675
Recorded/Played Tris ratio: 4.08
!TimeDemo Run 2 Finished.
Play Time: 64.68s, Average FPS: 23.19
Min FPS: 5.50 at frame 478, Max FPS: 34.03 at frame 102
Average Tri/Sec: -4785263, Tri/Frame: -206324
Recorded/Played Tris ratio: 4.05
!TimeDemo Run 3 Finished.
Play Time: 71.60s, Average FPS: 20.95
Min FPS: 3.87 at frame 473, Max FPS: 34.03 at frame 102
Average Tri/Sec: -4300051, Tri/Frame: -205256
Recorded/Played Tris ratio: 4.07
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Crysis CPU Benchmark #1

Here are the results from CPU benchmark #1. Same settings and the GPU benchmark and about the same 1-2 FPS improvement.

New (v1.2 and 169.44)
==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 41.61s, Average FPS: 36.05
Min FPS: 22.68 at frame 837, Max FPS: 51.96 at frame 342
Average Tri/Sec: -49459484, Tri/Frame: -1372004
Recorded/Played Tris ratio: -0.52
!TimeDemo Run 1 Finished.
Play Time: 35.69s, Average FPS: 42.03
Min FPS: 22.68 at frame 837, Max FPS: 55.96 at frame 100
Average Tri/Sec: -57293856, Tri/Frame: -1363148
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 36.24s, Average FPS: 41.39
Min FPS: 22.68 at frame 837, Max FPS: 56.02 at frame 102
Average Tri/Sec: -56125304, Tri/Frame: -1355871
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 35.99s, Average FPS: 41.68
Min FPS: 22.68 at frame 837, Max FPS: 56.02 at frame 102
Average Tri/Sec: -56813796, Tri/Frame: -1363019
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Old (v1.1 and 169.28)
==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 42.62s, Average FPS: 35.19
Min FPS: 19.99 at frame 822, Max FPS: 48.93 at frame 335
Average Tri/Sec: -48164704, Tri/Frame: -1368547
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 1 Finished.
Play Time: 37.05s, Average FPS: 40.48
Min FPS: 19.99 at frame 822, Max FPS: 48.97 at frame 372
Average Tri/Sec: -54835372, Tri/Frame: -1354615
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 37.05s, Average FPS: 40.49
Min FPS: 19.99 at frame 822, Max FPS: 49.96 at frame 93
Average Tri/Sec: -55253652, Tri/Frame: -1364712
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 36.93s, Average FPS: 40.62
Min FPS: 19.99 at frame 822, Max FPS: 49.97 at frame 143
Average Tri/Sec: -55146080, Tri/Frame: -1357689
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

New Crysis GPU Benchmarks

Here is the GPU benchmark with the new v1.2 patch and new Forceware 169.44 beta drivers. As you can see it is a slight improvement over v1.1 and Forceware 169.28.

It looks like I am getting about a 1-2 fps improvement over the old patch and drivers. In Crysis every frame counts and it's nice to see them moving in the right direction.

Again these tests are conducted on Windows XP 32 bit, 1920x1200 res, Q6600 @ 3.2 GHz, 4GB Patriot Viper @ 4-4-4-12-2T, and GTS 8800 512mb (G92) X 2 in SLI.

New (v1.2 and 169.44)
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 58.16s, Average FPS: 34.39
Min FPS: 20.83 at frame 138, Max FPS: 48.95 at frame 906
Average Tri/Sec: -22230674, Tri/Frame: -646482
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 49.48s, Average FPS: 40.42
Min FPS: 20.83 at frame 138, Max FPS: 50.96 at frame 92
Average Tri/Sec: -26059334, Tri/Frame: -644772
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 2 Finished.
Play Time: 49.01s, Average FPS: 40.81
Min FPS: 20.83 at frame 138, Max FPS: 52.96 at frame 89
Average Tri/Sec: -26296332, Tri/Frame: -644420
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 3 Finished.
Play Time: 48.79s, Average FPS: 40.99
Min FPS: 20.83 at frame 138, Max FPS: 52.96 at frame 89
Average Tri/Sec: -26408458, Tri/Frame: -644254
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Old (v1.1 and 169.28)
==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 61.60s, Average FPS: 32.47
Min FPS: 16.39 at frame 154, Max FPS: 42.97 at frame 924
Average Tri/Sec: -20968412, Tri/Frame: -645797
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 53.46s, Average FPS: 37.41
Min FPS: 16.39 at frame 154, Max FPS: 51.78 at frame 83
Average Tri/Sec: -24140702, Tri/Frame: -645223
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 2 Finished.
Play Time: 51.34s, Average FPS: 38.96
Min FPS: 16.39 at frame 154, Max FPS: 51.78 at frame 83
Average Tri/Sec: -25063880, Tri/Frame: -643366
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 3 Finished.
Play Time: 51.45s, Average FPS: 38.87
Min FPS: 16.39 at frame 154, Max FPS: 51.78 at frame 83
Average Tri/Sec: -25104210, Tri/Frame: -645802
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================

Friday, February 15, 2008


Here is my latest 3Dmark06 score! I switched to Patriot Viper memory with 4-4-4-12-2T timings as compared to the Corsair XMS2 with 5-5-5-18-2T timings.
Posted by Picasa

Thursday, January 24, 2008

F.E.A.R. Results


Here are the results of F.E.A.R. performance testing at 1920x1200 with 4X AA and all settings at maximum.
Posted by Picasa

Monday, January 21, 2008

Prime95: 6hrs 39min


Well we have reached a stable overclock vCore settings of 1.35000V. This has allowed the computer to run the Prime95 torture test for 6hrs 39mins. Yay!
Posted by Picasa

Sunday, January 20, 2008

3DMark06 Score of 17095!




How bout that for a 3DMark06 score.
Posted by Picasa

Saturday, January 19, 2008

More Crysis Screenshots





Here are some more screen shots for your enjoyment!
Posted by Picasa

Crysis Screen Shots




Here are some screen shots for your enjoyment!
Posted by Picasa

Crysis CPU #2 Benchmark

Again, same settings as the other two. 3.2GHz, 1920x1200 res, all settings on "high"

Running CPU benchmark 1
Results will depend on current system settings
Press any key to continue . . .
Running...
==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 127.52s)
!TimeDemo Run 0 Finished.
Play Time: 76.64s, Average FPS: 19.57
Min FPS: 5.50 at frame 478, Max FPS: 32.37 at frame 102
Average Tri/Sec: -3510228, Tri/Frame: -179346
Recorded/Played Tris ratio: 4.66
!TimeDemo Run 1 Finished.
Play Time: 64.52s, Average FPS: 23.25
Min FPS: 5.50 at frame 478, Max FPS: 32.71 at frame 91
Average Tri/Sec: -4758319, Tri/Frame: -204675
Recorded/Played Tris ratio: 4.08
!TimeDemo Run 2 Finished.
Play Time: 64.68s, Average FPS: 23.19
Min FPS: 5.50 at frame 478, Max FPS: 34.03 at frame 102
Average Tri/Sec: -4785263, Tri/Frame: -206324
Recorded/Played Tris ratio: 4.05
!TimeDemo Run 3 Finished.
Play Time: 71.60s, Average FPS: 20.95
Min FPS: 3.87 at frame 473, Max FPS: 34.03 at frame 102
Average Tri/Sec: -4300051, Tri/Frame: -205256
Recorded/Played Tris ratio: 4.07
TimeDemo Play Ended, (4 Runs Performed)
==============================================================
Press any key to continue . . .

Crysis CPU Benchmark

Here is the printout from the CPU benchmark. Also at 3.2GHz, Vcore=1.187v, 1920x1200 res with all settings on "high"

Running CPU benchmark 1
Results will depend on current system settings
Press any key to continue . . .
Running...

==============================================================
TimeDemo Play Started , (Total Frames: 1500, Recorded Time: 44.62s)
!TimeDemo Run 0 Finished.
Play Time: 42.62s, Average FPS: 35.19
Min FPS: 19.99 at frame 822, Max FPS: 48.93 at frame 335
Average Tri/Sec: -48164704, Tri/Frame: -1368547
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 1 Finished.
Play Time: 37.05s, Average FPS: 40.48
Min FPS: 19.99 at frame 822, Max FPS: 48.97 at frame 372
Average Tri/Sec: -54835372, Tri/Frame: -1354615
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 2 Finished.
Play Time: 37.05s, Average FPS: 40.49
Min FPS: 19.99 at frame 822, Max FPS: 49.96 at frame 93
Average Tri/Sec: -55253652, Tri/Frame: -1364712
Recorded/Played Tris ratio: -0.53
!TimeDemo Run 3 Finished.
Play Time: 36.93s, Average FPS: 40.62
Min FPS: 19.99 at frame 822, Max FPS: 49.97 at frame 143
Average Tri/Sec: -55146080, Tri/Frame: -1357689
Recorded/Played Tris ratio: -0.53
TimeDemo Play Ended, (4 Runs Performed)
==============================================================
Press any key to continue . . .

Crysis GPU Benchmark

Here is the print out from the Crysis GPU benchmark.

Running GPU benchmark 1
Results will depend on current system settings
Press any key to continue . . .
Running...

==============================================================
TimeDemo Play Started , (Total Frames: 2000, Recorded Time: 111.86s)
!TimeDemo Run 0 Finished.
Play Time: 61.60s, Average FPS: 32.47
Min FPS: 16.39 at frame 154, Max FPS: 42.97 at frame 924
Average Tri/Sec: -20968412, Tri/Frame: -645797
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 1 Finished.
Play Time: 53.46s, Average FPS: 37.41
Min FPS: 16.39 at frame 154, Max FPS: 51.78 at frame 83
Average Tri/Sec: -24140702, Tri/Frame: -645223
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 2 Finished.
Play Time: 51.34s, Average FPS: 38.96
Min FPS: 16.39 at frame 154, Max FPS: 51.78 at frame 83
Average Tri/Sec: -25063880, Tri/Frame: -643366
Recorded/Played Tris ratio: -1.42
!TimeDemo Run 3 Finished.
Play Time: 51.45s, Average FPS: 38.87
Min FPS: 16.39 at frame 154, Max FPS: 51.78 at frame 83
Average Tri/Sec: -25104210, Tri/Frame: -645802
Recorded/Played Tris ratio: -1.42
TimeDemo Play Ended, (4 Runs Performed)
==============================================================
Press any key to continue . . .

This was at a 3.2GHz overclock with Vcore set to 1.3187 and a 2 hour stable prime95 run.

Monday, January 14, 2008

Overclocking, SLI, and Crysis

Ok, so the first thing I did once I got everything working was fire up crysis. I have not played it much yet so I was really excited to see how it looked.

It looked like a slide show. 10fps. Boo, how come my kick ass Overclocked, SLIed, bitchin computer would not run the game it was freaking built for. Boooo!

Ok, so I did not give up there. I searched the nets and came up with some things to try and make it better.

Here is how I was running the game:

1920x1200 with all high settings = 10fps, slide show

1920x1200 with all medium settings = 20-30fps, somewhat playable

1.) Install the Crysis 1.1 patch. On its own this did not do very much, but it's important later.

2.) Install the 169.28 Forceware Beta Drivers. This doubled my FPS. I was amazed. I went from an average of 30 to an average of 60.

I am going to have to go back to "High" Settings and see how those look. Also, it doesn't seem like you have to do anything to the profile to enable SLI, even though it looks like it is off.

Anyways I'll post some screen shots of the overclocking and the crysis performance soon!

I am running a conservative overclock of 2.8GHz right now, all I did was turn the FSB on up. No messing with voltages or anything. I am going to try to get it to do 3.2 at some point, but where I am now is freaking fast.

Saturday, January 12, 2008

PSU in the eBay Version

For the eBay version of this computer I will be using the 800 watt version of the Tagan BZ. It is a high quality PSU that should be more than enough for powering a computer of this spec.

That is of course unless someone wants an upgrade:)

nForce 780 upgrade

Well, eVGA is offering an upgrade on recently purchased 680i motherboards. You can send in the MoBo and they will send you back a 780.

The main differences are support for 45nm Intel CPUs, support for three way SLI, and PCIe 2.0.

These could be major upgrade paths for me later on with this system so I am considering doing the upgrade. I guess I have about 30 days left to decide.


Here is the latest build. This is the Tagan BZ 1100. So far it runs completely stable in SLI mode with none of the problems I had before. More to come on Overclocking the CPU.
Posted by Picasa

Friday, January 11, 2008

What Now?

Well now I am running an 1100 watt Tagan BZ. I will have more info on it possibly tomorrow when I try to finalize the OC on this CPU.

This PSU is overkill for this system. I could have used the 800 watt version of it and been fine but it was a week moment and I went for the 1100 :)

I have been running crysis in SLI tonight and I am extremely CPU limited. MUST OC!!!

More details are coming!

jonnyguru.com

Visit http://www.jonnyguru.com/ for PSU info. It was helpful to me.

The Guilty Party

So as it turns out my PSU was starving GPU #2 and the CPU for amps when GPU #2 went into 3D mode. Why is this?

It all has to do with how the PSU is laid out electrically (needless to say this was something I should have paid more attention too when buying the PSU, but it was on sale at newegg.com and I didn't do as much research as I should have. What can I say?)

So here is how this OCZ GameXStream is laid out:

12v Rail 1: CPU #1
12v Rail 2: PCIe#2/CPU#2
12v Rail 3: Mobo / Acc.
12v Rail 4: PCI3#1

So, whats to wrong with that. Well, what they call CPU#2 is actually the second 4 pin cpu power plug. So if you have a motherboard with an 8 pin cpu power plug you will be using cpu 1 and cpu 2 on this power supply, which I was.

But wait, you say: each rail is rated for 18 amps. That should still be enough right?

Well yes and no. Each rail on it's own could hold an 18 amp load, but if you read the fine print on the PSU label you will see another power restriction. All of the 12v rails combined should never exceed 50 amps and there we reach the real problem.

OCZ reaches a 700 watt output by adding:
155 watts for the +3.3v and +5v rails
525 watts for the +12v rails
20 watts for -12v and +5Vsb

155 + 525 + 20 = 700 watts.

Ok I see that but there are still some things that don't add up. To reach 50 amps on the +12V rails you would need to be pushing 600 watts on the +12V rails, not 525. Oops, that must be some fuzzy math.

Ok so how many amps are you really getting if not 50amps?

525W / 12V = 43.75 amps

Oh boy! So how many of those did we need again?

CPU: 95 watts @ 12V = 7.9 amps
GPU #1: 146 watts @ 12v = 12.2 amps
GPU #2: 146 watts @ 12v = 12.2 amps

That leaves just
11.45 amps or 137.4 watts for the rest of the system, the motherboard, ram, DVD, hard drives, everything.

This math did not work out in my favor. The long story short is that I got an 1100 watt Tagan BZ psu and all the problems went away.

I would have never been able to overclock the cpu with the 700 watt PSU.

More SLI trouble

I enabled SLI as "alternate frame rendering 1" mode. This is important later.

Ok, so after all of the driver drama I enable SLI and loaded up Crysis. And I have so say I was not all that impressed with the performance and I had expected I would be with SLI.

There was also weird rolling greenish flicker over the whole screen. Needless to say I was less than impressed. I loaded FEAR and set it up for SLI and the same thing happened.

Same green rolling flicker. I could not capture it with screen capture. When I tried it always looked fine. It was almost like the refresh was too slow. Like when you look at a CRT television on a home video, where you can see the refresh.

Anyway this was a major bummer. Anytime I enabled SLI I got this weird flicker. There are a lot of people with this problem on the forums, but not may answers that I could find. Games ran find on 1 GPU but really poorly with 2.

I decided on a whim to try split frame rendering (where one card does the top, and another does the bottom of the screen) to see what that looked like. To my amazement the top half of the screen was perfect.

But the bottom half was all distorted and ugly and green. My second GPU was messing up!

When using alternate frame rendering, viewing the distortion every other frame looked like a flicker to the human eye.

So now I had tracked down the source of the problem, but why was my second GPU messing up? This was a major bummer.

I thought maybe I had a bad video card, so I pulled out the known good one, and swapped the second one to the first slot and loaded the system with just the one card. And it was fine. The computer worked perfectly with the one card in it.

So, I put the other card back in position 2. So now the cards were switched on the motherboard and the problem cropped back up again. Now I suspected the second GPU was being starved for power.

That shouldn't happen though, I've got an SLI ready PSU with 700 watts of power. Or do I?

What happend after I turned it on.

Ok, so the first time I hit the ON button it powered right up no problem.

LOL just kidding, I got a halt on floppy error. No problem, just a loose ribbon cable. I fixed that then tried to POST again.

This time no problem. So I set up the raid array and installed windows. Everything was going good. The PSU was kind of loud, but I could live with that.

So, once I got windows installed I loaded up the system drivers, updated windows, installed Crysis, and tried to enable SLI.

I say tried because I could not find the SLI control panel in the nTune application. It was not there. The computer could see both cards, and I was getting temp readings from both GPUs. So why couldn't I enable SLI?

I scoured the Internet forums looking for a solution. Everyone said the same thing, "reinstall the graphics drivers." Well I did this and it did not work.

After much trial and error I discovered where I went wrong. There is a very important, but almost never mentioned order to installing the system drivers on an nVidia system.

In order to get the nTune control panel to see everything you must install your apps in this order:

1.) Chipset Drivers and motherboard drivers
2.) Forceware Video Drivers
3.) nTune Application

If you install the forceware drivers first, then the chipset drivers, as I had done nTune will not properly recognize all of your hardware.

Built


Ok, so there it is. The first time I put it together. It lookes pretty good I think. I espcially like the two GPUs. Anyway keep on reading for what happened after I turned it on.
Posted by Picasa

OCZ PSU


Ok, so as I write this I am more than a little miffed at this PSU. Oh the drama it caused during the build. Just as a little spoiler the story did have a happy ending but this is not the PSU powering the computer as I write this post. More on that later.
Posted by Picasa

CPU Cooler


.
Here is the CPU cooler and the Arctic Silver 5. Never use any other kind of thermal grease!
Posted by Picasa

Memory


Here is the memory. Check out the cool head spreaders!
Posted by Picasa

DVD and Floppy


Here are the drives I am going to use. Pretty standard stuff. A DVD burner and a floppy (to load raid drivers with one time and them to never use again.) I like these drives that have the card reader build in. They cost a couple bucks more, but I think they are worth it.
Posted by Picasa

Monday, January 7, 2008

The Raptors


Ah yes, the 150gb Western Digital Raptors. This was one of the harder choices I had to make when I selected components for this system.

They use an old SATA interface, they are small (150gb), and they are expensive.

But... They spin at 10,000 rpms and are still the fastest SATA drives you can buy. And I have always wanted some and until now have not been able to get any.
Posted by Picasa

Graphics Package


Here is everything that came in the package with one of the video cards. It's nice that they include a copy of Crysis, but now I have three :(

Oh well, something else to eBay I guess. Anyway it seems like a pretty complete package. They include all the adapters that you are likely to need, though I probably won't be using any of them.
Posted by Picasa

2 GeForce 8800 GTS's


Oh yes! Count'em, 2 GeForce 8800GTS's. These are the G92 chip revision. They should use a little less power and put out a lot more polygons.

I guess we'll see when I get them installed.
Posted by Picasa

CPU SLACR


Of some importance when buying a Q6600 cpu is getting one with G0 stepping. You can see the SLACR engraved on the head spreader on this chip indicating G0 stepping.

In case anyone is trying to locate one like this, I bought this one from NewEgg.com in the retail box. It shipped out of New Jersey.
Posted by Picasa

The Q6600


The CPU arrived today. It is a Core 2 Quad Q6600, and I chose it because I believe it is the best value in CPUs right now. Some might argue that a Quad is overkill for a home PC because not enough applications are written to take advantage of that many cores.

But, because of its low price and high overclock ability I chose it anyway.
Posted by Picasa

Saturday, January 5, 2008

The 500 watt Antec PSU


This is the 500 watt psu that came with the Sonata III. It is not a bad psu at all, but it is not rated for SLI, and would be iffy for SLI and an Overclocked quad.

Anyway, it will be replaced with something that isn't here yet.
Posted by Picasa

Case accessories


Here are the accessories that come in the case. Again, this is pretty standard. The manual is a little sparse, but what do you expect? The case is pretty self explanatory.
Posted by Picasa