View Single Post
  #13 (permalink)  
Old April 24, 2008, 08:56 AM
PEIGuy PEIGuy is offline
Rookie
 
Join Date: Apr 2008
Location: Charlottetown
Posts: 8
Default

***UPDATE***

Well I have completed my upgrades and preformed some rough bench marks. Here is what I found out along the way...

Upgrade installation:



- Power supply was upgraded from the stock 305watt Dell to a 600watt OCZ StealthXStream. This went smooth with the 4 screw wholes lining up nicely. I really like this OCZ power supply, it is whisper quiet and has a mother load of connectors all nicely wrapped in mesh to help keep it organized.

- Futureshop was nice enough to "loan" me an EVGA 8800 GT and a BFG 8800 GTS for testing. I just couldn't get the GTS out of my mind; I had to see if I could justify it somehow. After testing I did return the cards to FS and then placed my order with NCIX.

I had an idea that I was going to have to modify the HS/F shroud to fit the GTS, but was unprepared for just how big that beast is. The GT on the other hand slipped right in see below for some pics.

BFG 8800 GTS



EVGA 8800 GT



To get the GTS in I would of had to slice the whole side out of the HS/F shroud as well as cut the first few fins off the chipset heat sink. When I seen how well the GT fit it made the decision even easier.

My reading led me to believe the GT ran hot and had me worried about the stock cooler just blowing all that hot air around the case. So I added a fan on the back side of the case to suck some of it out. In the end I don't think I had much to worry about because the Dell BTX case design actually works really well for keeping the rig cool'n'quiet.

Completed upgrade:




Now for the bench marks:



I used F.E.A.R, Crysis and 3DMark06 for testing, and RivaTuner for OC'ing the cards. These provide a rough gauge more than anything.

XFX 8600GT OC'ed to 650 CC and 450MC -VS- EVGA 8800GT with 600CC and 900MC (stock)

F.E.A.R (built in test)

8600GT after 5 test runs:

- 1280x960 res with max settings including 4x AA

FPS = 12 min, 22 Avg, 28 Max

66% below 25
34% between 25 and 40
0% over 40

8800GT after 5 test runs:

FPS = 22min, 26avg, 31max

48% below 22
52% between 25 & 40
0% over 40

CRYSIS (first good fire fight on beach with Fraps)

8600GT

- 1024x768 res with Medium settings no AA

FPS = 11 min, 21 Avg, 31 High


8800GT

- 1680x1050 res with Medium settings no AA

FPS = 10 min, 25 Avg, 37 High



3DMark06 (basic)

8600GT

- 1280x1024 res with stock settings
1st run 3DMarks = 4354
2nd run 3DMarks = 4372

8800GT

- 1280x1024 res with stock settings
1st run: 3DMarks = 8913
2nd run: 3DMarks = 8993

- 1280x1024 res with 700CC and 1000MC
1st run: 3DMarks = 9229
2nd run: 3DMarks = 9260


Note 1: Overclocking the 8800GT to 700CC and 1000MC had no effet on FPS for either F.E.A.R or CRYSIS, but did in fact give a minor boost to 3DMark for some reason?


Note 2: The hotest I recorded the 8600GT at was 82 degrees celcius and 72 degrees for the 8800GT. And that was with the 8600GT fan set to a 100% while the 8800GT is set to auto. The 8800GT idles at 54 degrees.


Conclusoin:

I am really glad I didn't go ahead and mod my HS/Fan shroud to fit the 8800 GTS. Overclocking the 8800 GT with zero net result in FPS in both F.E.A.R and CRYSIS really put it in black and white for me that my CPU simply can't feed the beast. The only spot overclocking the 8800 GT made a diffence was in 3DMark06 taking the score from 8993 to 9260.

Having said that I am very pleased with the jump my games took going from the 8600GT to the 8800GT. It has been worth the time and cash.

Lastly going through this little exercise really put in prespective how much work it is to accurately test/benchmark graphics cards. Simply put it takes time and care to do it right.

Last edited by PEIGuy; April 24, 2008 at 10:23 AM. Reason: Major edits for many reasons ;)
Reply With Quote