As I mentioned in the previous article, the advantage of the retail Radeon 9800 Pro over the BTO 9800 Pro card is the fact only the retail model can use the ATI displays 3D/GL Overrides feature. But the downside is that the retail card does not have an ADC port (for those that own or plan to buy an ADC display). A $99 adapter would be required for the retail card. (The BTO card can also drive two Digital Displays, where the retail card has one analog and one Digital/DVI port, but it also has Video/TV out.) The 9800 Pro retail FAQ has screenshots and info on the 3D/GL overrides feature, which allows you to enable FSAA, Anisotropic filtering even in games that do not normally support those features. The BTO card is an 8x AGP version, where the retail is 4x AGP, but as I suspected, that's not a significant factor in any of the tests I ran (at least with current apps/games). As you'll see from some of the test results, the retail card was a tiny bit faster than the OEM card at higher resolutions in some tests. The rumor that the BTO/OEM card may be slightly lower clocked than the retail card may be a reason why.

BTO/OEM vs Retail 9800 Pro Summary:

• No real significant performance differences in the cards (although the OEM card is clocked slightly lower)
• Only the Retail 9800 can use the ATI displays control panel 3D/GL Overrides and Versavision (display rotation) features. (Update: Long after this article was posted, ATI later revised the Displays Control panel app so that 3D overrides could be used on some OEM Radeons, although versavision isn't supported with OEM cards, due to firmware versions IIRC)
• Retail 9800 has Video Out/TV Output port
  (w/o needing the Apple $19 DVI->Video adapter for G5s)
• Retail 9800 can be used in older G4/AGP towers (BTO 9800 card requires an AGP Pro slot)
• BTO 9800 card has ADC port (for Apple LCD use w/o adapters)
• BTO 9800 does not require using Aux. Power Supply Cable
• BTO 9800 can drive two digital displays (DVI + ADC or two DVI displays)
• BTO 9800 not sold separately (yet at least), only available with G5 system purchase

(OEM has DVI-I and ADC Ports, notice ADC stub/AGP Pro slot connector)

(Retail has VGA, Video Out and DVI-I Ports, Note Pwr Supply connector)

For notes on installing the retail 9800 Pro in a G5 (aux power cable connection), see that section of the previous article.

One thing I have noticed - the fan on this particular BTO/OEM 9800 Pro is louder than the one on the retail card I have. This may not be typical but this one is annoying (I can stop the fan with my finger and the noise stops - so it's clearly from the card's fan.) I'm going to have to replace that BTO card's fan with a quieter one, perhaps an aftermarket cooler/fan as the noise frequency/tone is getting on my nerves. (Update: I replaced the noisy OEM Heatsink/fan with the expensive but very quiet Verax G03 Cooler)

G5 Test System Details: (All OS X updates as of Oct. 20th, 2003 applied before testing)

• PowerMac G5 Dual 2GHz OS 10.2.8 (G5), boot rom 5.0.2f2, 1.5GB RAM (CL3), 160GB Seagtate hard drive, Superdrive (Pioneer DVR-106D), Airport Extreme card (enabled/active for all tests)
• Retail Radeon 9800 Pro card ROM: 113-A07525-114
• BTO/OEM Radeon 9800 Pro card ROM: 113-A14401-116

I'm not sure if current retail 9800 Pro cards have a later ROM version than my early sample, but I have seen reports of a "-117" BTO/OEM ROM version. (Not sure what is changed in 117 vs 116).

This game is very CPU/Bandwidth bound (i.e. often delivering similar FPS rates across resolutions with high end cards). If you saw my original page of G5 tests (vs MDD Dual 1.25GHz) you know that the Dual G5 delivers much higher scores than previous Macs, even with the same 9800 Pro (retail) card.

It's old, but Quake3 is still popular and one of the few on the Mac that seems to scale well. All tests used the latest Quake3 1.32 "G4" version. Sound was enabled for all tests but music volume was set to 0 and all game options were on. First a test with Quake3's "high quality" settings + high geometric detail and maximum texture quality. (r_smp set to 1 to use dual CPUs, and my typical "s_chunksize 4096" vs default 2048 was used..)

(Note - the standard "High Quality" settings w/o max texture quality/high geometric detail would show a bit higher rates than the above "max" settings - for example with std "high quality" only settings and no config file tweaks I get 340FPS at 1024x768 and 223.2 FPS at 1600x1200.)

Although with a high end system and graphics card, I'm more interested in high quality settings performance, I also ran a test with the "Fastest" setting (low res/16bit/ugly), just to see what speeds would be seen on the G5 vs G4. (Use of custom configs like boli's also can boost framerates even higher, at the expense of image quality.)

Note: after a reader request, I ran 1600x1200 tests on the Dual G5/9800 retail combo with sound disabled and standard "High Quality" settings (which have medium Geometric detail and Texture quality slider down from Max. The "Max Quality" tests I ran had HQ + high Geometric detail and Texture quality slider set to max, which takes a bit of a performance hit). The dual G5 2GHz w/retail 9800 pro card (sound disabled and std HQ settings) delivered 227.4 FPS at 1600x1200 and 400.1 FPS at 640x480 under 10.2.7.

All UT2003 tests used max image quality settings and the benchmark procedure mentioned on the UT2003 demo tests entry page.
Since the game is already a CPU hog, I modified the INI file to set Audio channels to 4. (The flyby test defaults to no audio, but the botmatch test does have sound enabled.) I used the Asbestos benchmark, since it was less CPU bound than the outdoor Antalus level (in both flyby and botmatch).

Note: Since only the retail ATI 9800 Pro can use the ATI 4.1 displays control panel 3D/GL overrides feature (for selecting FSAA, Anisotropic filtering, etc.) I used edited UT2003.ini files to enable FSAA and Anisotropic filtering using the tips noted months ago in the Game section of my xlr8yourmac.com FAQ. (No longer online)

I used the same RTCW Multiplayer demo tested used in CPU upgrade reviews, sent by a reader. (Demo from button in the sidebar at http://ultimatedataserver.de.vu/.) Note: With RTCW v1.41, you have to change the "59" in the demo filename to "60". Once the demos folder is installed, just open the console using ~ and type Timedemo 1 and Demo demo0001.) All tests used High Quality game settings, audio enabled, but music volume set to 0.

Again as resolutions rise, the retail card was a tiny bit faster.

As I mentioned in the previous article with G5 tests, I'd seen some unexpectedly low FPS rates (with the instantenous FPS counter enabled) in single player RTCW even with fast G4 systems. I keep a saved game to test for this (from one of the first levels, where you're in the room that has the holes in the floor with solders in the lower room). Standing in the corner of that room even without any action happening shows dips into the 20s, with averages in the 38-45FPS range (at 1024x768/HQ). With the G5, the FPS meter was usually pegged at 80-90FPS and never dropped anywhere near as low as it did with previous Macs.

It's not a great test of video card performance typically but it's still a popular game. Unreal Tournment X PR3 was set to Maximum texture/skin detail and MinFPS = 0, audio on and music set to min volume. I ran tests at 4x FSAA, 1024x768 mode (highest possible with super-sampling, although the 9800 supports Multi-sampling some older cards don't) and also at 1600x1200 (w/o FSAA). I used my own recorded demo (actual FPS rates during typical play were higher).

I ran Lightwave 3D 7.5c on the Dual G5 and ran the "Hummer" scene preview playback test. Since this test uses a stopwatch there's some variance in start/stop times so consider times within 1/2 second to be literally indentical.

Playback Hummer Preview (full size)
• Dual G5 2GHz/9800 Retail (10.2.8): 23.6 sec.
• Dual G5 2GHz/9800 BTO (10.2.8): 23.1 sec.

As a FYI - a Dual G4 1.25GHz with 9800 retail card (10.2.6) took 46.3 seconds for this test.

With the desktop set to 1600x1200, millions colors I used AppleWorks 6.2.7 to test scrolling times from top to bottom of a 100 page, multi-column newsletter with images. As usual, in the same system, the differences with different graphics cards is small (within the margin for error of starting/stopping a stopwatch). I didn't bother to graph the results since both the retail and BTO 9800 cards had literally the same time (10.2 seconds).
As shown in past articles/reviews, 2D performance isn't really graphics card bound with a modern graphics card. (I.E. using the same system, most modern cards will deliver similar performance in tests like this.)

I used Maxon's Cinebench 2003 benchmark G5 beta version with the desktop set to 1600x1200, millions colors. I've included both the OpenGL scores as well as the Framerates for the two 3D scene fly-by tests. (I didn't graph the render score since this article is on graphics cards and used the same system. As a FYI, the Dual G5 2GHz render score ws 514/503 for with the BTO/Retail 9800 card installed.)

I also graphed the detailed results from CineBench regarding the Hardware OpenGL framerates for the Scene1 and Scene2 fly-bys.

(I didn't include Xbench tests since its graphics card tests are fairly primitive and don't really stress a graphics card enough to show any significant difference in these two 9800 cards.)

Other G5 Related Articles: See the Mac Mods and Upgrades page.

= Other Site Topic Areas =
Mac Upgrades/Mods | Storage | Video | Audio/Home Theater | macOS Updates/Tips/Wi-Fi