Contribute
Register

confused about 4xx card performance

Status
Not open for further replies.
Joined
Oct 9, 2010
Messages
49
Motherboard
Ga-Z170x-UD5 TH
CPU
6700k 4.8ghz
Graphics
GTX 980 ti FTW
Mac
  1. Mac Pro
Classic Mac
  1. 0
Mobile Phone
  1. 0
ok so i have been watching the 4xx thread hoping that someone would figure out how to get a 5xx card running when people started posting cinebench scores. some are in the low 20's.

i downloaded it and ran it and this is what i got.
00a32612eaaa11a6e2f0976bba69e2f9.png


the orange result is after installing the new drivers then modifying the ktext with the device ID
the brown result is before nvidia drivers were installed and no ktext modification.

why would anyone spend double on a card that performs worse than a gtx-260 in a hackintosh?

is this test innaccurate because of the processor speed?
 
Looks about right to me. I believe I read that people were getting 26-28 fps with those cards. Remember, it's just a number, and doesn't tell the whole story.

How does it perform elsewhere? Editing, gaming, etc?
 
I have two OSx86 systems, the first running 10.5.6 with a GTX260 and a second one that I just got up and running with 10.6.5 and a GTX460.

I ran some bench marks on both systems and surprisingly the GTX260 outperforms the GTX460 in both Xbench and Cinebench. Here are the scores:

GTX260 - Xbench:
Quartz Graphics Test - 299.55
OpenGL Test - 204.80

GTX460 - Xbench:
Quartz Graphics Test - 262.01
OpenGL Test - 172.01

GTX260 - Cinebench
OpenGL - 23.97 fps

GTX460 - Cinebench
OpenGL - 22.03
 
XFX 285 GTX -- Cinebench ~24fps
EVGA 470 GTX -- Cinebench ~24fps -- complete system freeze when playing WoW

Others have also reported hack system freezing when playing SC2.

The nVidia Fermi drivers does not seem stable enough yet. I returned the Fermi card, and maybe will revisit it when drivers are stable enough to play WoW and SC2.
 
I'm confused too. isn't GTX260 supposed to be much better than my GTS 250???
cinebench_OpenGL_4.2Ghz_gts250.png
 
Heyyo,

It might be your chameleon RC4/5, nullcpupowermanagement or smbios.plist MacPro3,1 or other!!!

I've been messing around.. check this!

Easybeast (defaults with Chameleon RC4 and NullCPUPowerManagement):
easybeast.png


EasyBeast with Chameleon RC5 r643 and NullCPUPowerManagement:
easybeastchameleonrc5nu.png


EasyBeast with Chameleon RC5 r643
easybeastchameleonrc5.png


Needless to say.. I was shocked by the results. Try using just norm easybeast with GE=Yes.. it seems odd to me that not having NullCPUPowerManagement causes a big framerate hit (and some stuttering).

Another interesting one was with smbios.plist set to anything other than MacPro 3,1 causes my framerates to be stuck at a maximum of 30F/S... i5/i7 doesn't matter what I pick there.

Btw for a decent preformance boost? just like Windows on CSS and TF2 the multicore rendering doesn't enable all multi threaded rendering capabilities.. so set this as your launch options for optimal netcode and preformance out of your cpu:

Code:
-heapsize 2097152 -novid +rate 30000 +cl_cmdrate 101 +cl_updaterate 101 +cl_interp_ratio 1.0 +mat_queue_mode 2 +r_threaded_renderables 1 +r_threaded_particles 1

My system build:
Processor:
Intel(R) Core(TM) i7 CPU 920 @ ~3.6GHz (8CPUs) (Using Coolit Domino ALC)
Memory:
8GB (4x2GB) DDR3 1066MHz OCZ Platinum CL 8-8-8-24 (Triple Channel)
Hard Drive:
4 Seagate 320GB SATAII HDD's in RAID 0
Video Card:
Zotac's NVIDIA GeForce GTX 460
Monitor:
LG 22" Widescreen L226WTQ(Digital)
Sound Card:
Asus Xonar DX (PCI-E)
Speakers/Headphones:
Antec Lansing 2.1 Speakers
Keyboard:
Logitech G15 Keyboard
Mouse:
Razer DeathAdder Mouse
Mouse Surface:
Razer Destructor
Operating System:
Windows 7 Ultimate 64-bit (6.1, Build 7600) (7600.win7_rtm.090713-1255)
Motherboard:
Asus P6T
Computer Case:
Thermaltake Armor (Silver)

[EDIT1]
Btw game settings during my testing...


Uploaded with ImageShack.us


[EDIT2]
and YES... this is directly comparable preformance to Windows 7 64bit in CSS with EasyBeast Chameleon RC5 with NullCPUPowerManagement cause I get roughly the same framerate in Win7.. Maybe later I'll try messing around with the dsdt.aml I found online a while back.. see if that affects preformance with and without NullCPUPowerManagement...

[EDIT3]
K... same results here... so the next two are with the dsdt.aml for my Asus P6T mobo from http://www.kexts.com

UserBeast ChameleonRC5 with NullCPUpmgnt
userbeastchameleonrc5nu.png


UserBeast ChameleonRC5
userbeastchameleonrc5.png


Hmm.. looks like I'm better off keeping the nullCPUpmgnt.. I also noticed stuttering issues quite a bit when I first shot my gun and stuff like that when I loaded into maps without the NullCPUpmgnt... so yeah, dunno why ChameleonRC5 and this dsdt.aml seem to have problems.. maybe they aren't detecting how many cores I have even though my smbios.plist is a MacPro 3,1 i7???
 
Consider yourself lucky, I get about 8,75 in Cinebench :lol:

Just some numbers

Cinebench Open GL

Mac 8.75 - Win 64 39.07

Cinebench CPU

Mac 1.9 - Win 64 4.97

Something needs work

WITH NullCpu :

Cinebench OpenGL : 24,79 CPU : 4,91

No worldbeaters, but a bit more respectable.

Here I was thinking I wouldn't need NullPower with a DSDT.

Maybe I should try and make an aml for the F15 bios revision of my mobo. OR there's a bios setting wrong somewhere.
 
Heyyo,

Heh I guess the nullcpumgnt holds true.. it's gotta be a bug in the dsdt.aml we have or something chameleon rc5's power management detection???

One other thing we must consider is the 400 series are not officially supported where-as Geforce gtx285 is at least supported so it probably got some of the patching love that graphics cards got before when people complained about low frame-rates in steam games... so I dunno if there's a way to truly "fix" our performance.. but at least be happy we can use our darn video-cards at least. :p
 
With my ASUS ENGTX285 :

Capture%20dcran%202010-12-04%20%2001.09.36.png



Have a big difference between GTX460/470 and my ENGTX285 on Mac OS X 10.6.5 ?
 
Status
Not open for further replies.
Back
Top