you should use it with vsinc on, i suppose you don't have a monitor @ 250 Hz+
Announcement
Collapse
No announcement yet.
Triple Buffering in 3.7s....Why?
Collapse
X
-
I know you should use it with V-sync on, that's the whole point of using TB. But I'm questioning the results of what has been posted.
EDIT: Sorry, thought that was directed at me.Last edited by BuddMan; 10 September 2003, 13:20.System Specs:
Gigabyte 8INXP - Pentium 4 2.8@3.4 - 1GB Corsair 3200 XMS - Enermax 550W PSU - 2 80GB WDs 8MB cache in RAID 0 array - 36GB Seagate 15.3K SCSI boot drive - ATI AIW 9700 - M-Audio Revolution - 16x Pioneer DVD slot load - Lite-On 48x24x48x CD-RW - Logitech MX700 - Koolance PC2-601BW case - Cambridge MegaWorks 550s - Mitsubishi 2070SB 22" CRT
Our Father, who 0wnz heaven, j00 r0ck!
May all 0ur base someday be belong to you!
Give us this day our warez, mp3z, and pr0n through a phat pipe.
And cut us some slack when we act like n00b lamerz,
just as we teach n00bz when they act lame on us.
For j00 0wn r00t on all our b0x3s 4ever and ever, 4m3n.
Comment
-
People have reported a more consistant fps rate with tb on. Thats over at rage3d anyway.
Comment
-
...i suppose you don't have a monitor @ 250 Hz+
With vsync always on in the drivers (monitor refresh at 100Hz):
TB on-100.2 fps
TB off-99.8 fps
Comment
-
@Admiral - you probably need to use a more graphics intensive program. Try the X2 demo (if you can _really_ force the settings you want in the drivers). The TB benefits should come into play when the graphics card can keep up on average, but not necessarily on every frame.
I'd be curious to see the benefits on 3 types of games: those that the GPU has no hope of keeping up with (X2), those that the GPU is able to create roughly the same frame rate as the monitor refresh (RTCW/ Wolf ET?), and those that the GPU is hopelessly overpowered for (Quake2?). (these cases can probably be simulated by varying the resolution of a single game)
- Steve
Comment
-
Originally posted by Admiral
My bad, the above test was done using default off for vsync in the drivers and enabling and disabling sync every frame in the game.
With vsync always on in the drivers (monitor refresh at 100Hz):
TB on-100.2 fps
TB off-99.8 fpsSystem Specs:
Gigabyte 8INXP - Pentium 4 2.8@3.4 - 1GB Corsair 3200 XMS - Enermax 550W PSU - 2 80GB WDs 8MB cache in RAID 0 array - 36GB Seagate 15.3K SCSI boot drive - ATI AIW 9700 - M-Audio Revolution - 16x Pioneer DVD slot load - Lite-On 48x24x48x CD-RW - Logitech MX700 - Koolance PC2-601BW case - Cambridge MegaWorks 550s - Mitsubishi 2070SB 22" CRT
Our Father, who 0wnz heaven, j00 r0ck!
May all 0ur base someday be belong to you!
Give us this day our warez, mp3z, and pr0n through a phat pipe.
And cut us some slack when we act like n00b lamerz,
just as we teach n00bz when they act lame on us.
For j00 0wn r00t on all our b0x3s 4ever and ever, 4m3n.
Comment
-
I also have STV: Elite Force and Elite Force 2 installed.
Guess I should choose me a spot in Elite Force 2. In the first part of the game, on the Borg sphere where fps tend to be the lowest in the game.
I can also install MDK 2 or run a benchmark for UT2003 in OpenGL mode.
edit: not much luck, MDK 2 test settings also shows a small difference, 1 or 2 fps, depending on the resolution.
UT2003- though the game is set to run (and runs in) OGL the benchmark seems to still take D3D. If someone could tell me what file to change ?
EF 2- I chose Borg Sphere level 1b, where Voyager is shown in the dampening field inside the sphere... can't tell the difference.
I'll have to download the X2 demo.Last edited by Admiral; 10 September 2003, 15:21.
Comment
-
Ok, well here are my results with UT 2003. Resolution was 1280x1024 @ 100Hz using Omega's 3.7s, hacked version. I benchmarked with FRAPS, and this was the Suntemple Botmatch:
2003-09-10 22:11:23 - UT2003 - VSYNC ON TB ON
Frames: 1827 - Time: 34469ms - Avg: 53.004 - Min: 39 - Max: 97
2003-09-10 22:12:32 - UT2003 - VSYNC ON TB OFF
Frames: 1826 - Time: 35156ms - Avg: 51.939 - Min: 36 - Max: 95
2003-09-10 22:23:54 - UT2003 - VSYNC ON TB ON 4XAA
Frames: 1832 - Time: 38922ms - Avg: 47.068 - Min: 28 - Max: 55
2003-09-10 22:25:40 - UT2003 - VSYNC ON TB OFF 4XAA
Frames: 1832 - Time: 38532ms - Avg: 47.544 - Min: 37 - Max: 54
I did the tests countless times and every time the result was pretty much the same. So given this, does it look like TB is working as it should be?
EDIT: And yes, I ran it using the OpenGL renderer and AF was on app preference.Last edited by BuddMan; 10 September 2003, 20:36.System Specs:
Gigabyte 8INXP - Pentium 4 2.8@3.4 - 1GB Corsair 3200 XMS - Enermax 550W PSU - 2 80GB WDs 8MB cache in RAID 0 array - 36GB Seagate 15.3K SCSI boot drive - ATI AIW 9700 - M-Audio Revolution - 16x Pioneer DVD slot load - Lite-On 48x24x48x CD-RW - Logitech MX700 - Koolance PC2-601BW case - Cambridge MegaWorks 550s - Mitsubishi 2070SB 22" CRT
Our Father, who 0wnz heaven, j00 r0ck!
May all 0ur base someday be belong to you!
Give us this day our warez, mp3z, and pr0n through a phat pipe.
And cut us some slack when we act like n00b lamerz,
just as we teach n00bz when they act lame on us.
For j00 0wn r00t on all our b0x3s 4ever and ever, 4m3n.
Comment
-
Ok, well if anyone else has a 9700 pro I'd like to see their results with TB. If TB doesn't work with AA on with the 9x00 series then more people should be aware of it. But my results alone can't really conclude anything yet. I'll have to sit down this weekend and install the normal 3.7s and see if that changes anything and benchmark a few more games such as Serious Sam 2. I'll post results when I have them.System Specs:
Gigabyte 8INXP - Pentium 4 2.8@3.4 - 1GB Corsair 3200 XMS - Enermax 550W PSU - 2 80GB WDs 8MB cache in RAID 0 array - 36GB Seagate 15.3K SCSI boot drive - ATI AIW 9700 - M-Audio Revolution - 16x Pioneer DVD slot load - Lite-On 48x24x48x CD-RW - Logitech MX700 - Koolance PC2-601BW case - Cambridge MegaWorks 550s - Mitsubishi 2070SB 22" CRT
Our Father, who 0wnz heaven, j00 r0ck!
May all 0ur base someday be belong to you!
Give us this day our warez, mp3z, and pr0n through a phat pipe.
And cut us some slack when we act like n00b lamerz,
just as we teach n00bz when they act lame on us.
For j00 0wn r00t on all our b0x3s 4ever and ever, 4m3n.
Comment
Comment