Confirmation of Bottleneck

[shadow187]

Comp1:
Q8200
4GB DDR3 1066
WinVista64
HD4650

Comp2:
E1400
2GB DDR2 667
WinVista32
HD4650

Comp1 (team fortress 2) @ 1680x1050 Vhigh Settings
40-70FPS

Comp2 (Team Fortress 2) @ 1280x1024 Low Settings
20-70FPS

During intense firefights, the frames drop horribly. The frames also drop when I look at a bigger area of a level.

Bottleneck confirmation, y/n?

 

[capt_taco]

For Comp1, yes, the video card is way underpowered compared to the rest of the system. You'd do a lot better if you got at least a 4800 series card.

For Comp2, there's no bottleneck; the whole computer just sucks.

Although since the second machine looks like it's getting too far out of date to be worth saving ... you might consider just cannibalizing the 4650 out of it and crossfiring both 4650s in the first machine. That way you might be able to squeeze out about the same performance as a 4850 without spending any money.

 

[Kari]

You should really overclock those CPUs (if you can).

 

[shadow187]

I don't need a powerful video card. This one plays all my games at max, and I am happy. Please just answer my question.
-Taco

I was thinking about it. If I go an extra 2-300mhz, it should give me about 10 frames or so. But I don't know why the frames are so low minimally and high maximally.
-Kari

 

[capt_taco]

I don't need a powerful video card. This one plays all my games at max, and I am happy. Please just answer my question.
-Taco

I just did.

 

[shadow187]

Then explain why a 7600GT and an HD4650 are getting the same frames.

 

[capt_taco]

What's this about a 7600GT now? Was the HD4650 in the second machine a typo?

 

[shadow187]

No. Two different cards. Same compy, same frames (though I haven't put the 7600GT in Comp: 1).

 

[notty22]

Then explain why a 7600GT and an HD4650 are getting the same frames.

With this card your not stating what system but your systems are both bottlenecking that game. those fps swings min-max---What counts is AVG.
On scenes with not much on screen/foreground maybe no smoke,no bullets, no clouds, no other players your cpu is fast enough to feed the data to the video card which in ideal conditions for a given resolution will get 70 fps. Move to a scene with more data to be processed your cpu can't feed it fast enough to your video card. Thats why same video card different cpu's same max fps. but not avg, by your abysmal low fps in some places.

With that game a slightly better video card, say a 4850 with a min of 3gh dua/ or quad would probably avg 70fps at 1650 resolution.
edit /c comp 1 would probably give you 57-70 instead of 40-70.

 

[capt_taco]

In that case, it sounds like the CPU is bottlenecking the second machine.

On the first machine, possibly the 4650 gets maxed out during the intense sequences. I wouldn't doubt it with the higher resolution and higher settings.

Have you tried running them at the same (lower) resolution and settings? In that case, I would bet the second machine slows down and the first doesn't. That would be another way to confirm a CPU bottleneck for sure.

 

[shadow187]

Already did. All high settings on Machine 2 gives me the same frames as all low. And I already guessed it was a bottleneck because I'm running a major resolution difference.

Ahh well, cherio. I'm off to overclock an nVidia nForce 610i.

 

[geofelt]

2gb of Ram may be the culprit in the minimum FPS arena for machine 2.
Read this study:
http://www.corsair.com/_appnotes/AN804_Gaming_Performance_Analysis.pdf

 

[shadow187]

OC'd to 2.4Ghz, minimum FPS went up 10, max went up 10. Going to push 2.7 tomorrow. Hope this machine can handle it!