Jump to content

Performance lacking (E7500 3.0, 9800 GT)


droodyf
 Share

12 posts in this topic

Recommended Posts

Hi all,

 

I've done a Lifehacker-guide style install, retail Snow Leopard with the Chameleon boot loader and that's all gone fine.

 

I'm using an E7500 (3.0ghz dual core) and a 9800 GT, and I have a suspicion I still have something to 'fix' with the install before its performing properly. Basically I'm being outperformed by my Mac Mini by a just-noticable amount.

 

The Xbench scores are all over the place, the macmini essentially beats the hackintosh in every area except scores, whereas in actual usage the two machines feel very similar, but the Mini has a slightly quicker feel to the interface and scrolling around in bigger windows.

 

I don't know whether this is simply because I'm running in higher res on the hackintosh, or whether something just isn't optimised as well as the Mini. Right now the mini is at 1680x1050 and the hackintosh at 1920x1200.

 

Xbench and OpenGL results below (1280x960x32) for OpenGL as its the first resolution I can run the same on both computers)

 

Is there anything obviously wrong that I've missed? This is the first hackintosh build I've done and I'm not 100% whether there's much post-install configuration to do.

post-492793-1254244472_thumb.png

post-492793-1254244481_thumb.png

post-492793-1254244484_thumb.png

Link to comment
Share on other sites

?????

Comparing OpenGLExtensionsviewer Bench values i think right pic shows GT9800 , left pic MacMini.

Cant see that the MacMini is "outperforming" the 9800GT !!

 

For better gpu compareage set also (beside FULL SCREEN) Multisampling to 4* and Benchmark - GPU has much more to do, less CPU dependend values.

Forget XBench for OpenGL benches - even an GMA950 is there faster than an ATI 2600 XT - unreal bench !

 

My 9600 GT (1280x960, 60 Hz (more HZ give a bit higher bench values!))

screenshoot attached (Fullscreen + Multi*4 + Benchmark) , your 9800 GT should perform up to 20% faster than mine.

 

 

To the rest of XBench (CPU, Thread, User Interface - all three usable for benching, but geekbench is also much better) :

Indeed your hackintosh is running slow .

Have you checked if your CPU is running full speed (CPU-X) or do you have any speedstep installed - Xbench CPU & Co values are always bad if speedstep is aktive and throtthles.

 

 

 

Use geekbench, also much better (newer) CPU + Mem Test.

Here my Values: E7400 , at 3.0 GHZ (by FSB+ to 333, Multi- to 9*) , 1066 DDR2 RAM

http://browse.geekbench.ca/geekbench2/view/152423

 

 

Compare especially Memory Speed - sometimes your memory speed may be bad by wrong bios settings (no OS X problem, set it up in the bios)

Bild_364.jpg

Link to comment
Share on other sites

Oh sorry, yes, I meant to say with OpenGL testing it was only the 2.1 mark that the Hackintosh lost, which was surprising given how fast it was in the others.

 

Thanks for the your suggestions about clock speed / ram timings - I'll get these checked out and see if I find anything. So far I've used the motherboard auto discovery results, which initially put my CPU at 2.13 ghz (8x multipler) but I could set it back to 11x to get 2.93 without problem, so maybe its got the ram timings all wrong as well.

 

I'll get Geekbench and try that now!

Link to comment
Share on other sites

Yes geekbench is much better bench since OS X 10.5 - Xbench never been updated since 10.3 time !!!!

 

My geekbench values shown are 64 Bit geekbench version , which are a bit higher than the 32 Bit geekbench, i remember to had around 3900 geekbench 33 Bit (to 4700 64 Bit) if you compare to mine. 10.6 geekbench near same / no major differences.

Link to comment
Share on other sites

Geekbench (32 bit):

Mac Mini 2917 v. Hackintosh 2500

 

CPU and Memory timings are correct according to BIOS / CPU-X.

 

The only thing I see is that the Mini has DDR3-1066 while the Hackintosh is DDR2-800 (The EP45-Extreme does not support DDR3). If this difference in speed is purely down to the ram difference, I can only conclude that ram is pretty important to OS X !

 

I don't know how tolerant OSX is to overclocking, but I will see if I can increase the timings a little, or try a faster DDR2 type :)

Link to comment
Share on other sites

Geekbench (32 bit):

Mac Mini 2917 v. Hackintosh 2500

 

CPU and Memory timings are correct according to BIOS / CPU-X.

 

The only thing I see is that the Mini has DDR3-1066 while the Hackintosh is DDR2-800 (The EP45-Extreme does not support DDR3). If this difference in speed is purely down to the ram difference, I can only conclude that ram is pretty important to OS X !

 

I don't know how tolerant OSX is to overclocking, but I will see if I can increase the timings a little, or try a faster DDR2 type :(

 

What exact are your Memory Geekbench Values of Hackintosh vs MacMini ?

My C2D 3.0 GHz, P35 Chipset, DDR2-1066 get:

Memory Memory performance 3500

Stream Memory bandwidth performance 2700

 

Your Mainboard with P45 Chipset is very good - but you dont need really DDR3 RAM BUT!!! why are you using 800 MHz RAM ?

Your CPU must be near mine but something is wrong - Memoryspeed or something other like background tools running ? speedstep voodoo running ?

Link to comment
Share on other sites

Hi Mitch,

 

I got the Geekbench figures for you:

 

Mac Mini:

Memory Performance - 2464

Stream Memory Performance - 1822

Integer - 2304

Floating Point - 4127

 

Hackintosh:

Memory Performance - 2168

Stream Memory Performance - 2119

Integer - 1976

Floating Point - 2820

 

I've put the full details up here:

Mac Mini and Hackintosh

 

I'm using DDR2 800 instead of 1066 because I (mistakingly) read that DDR2 800 was the most compatible, so I thought it would be a better choice.

 

I'm not sure what else is wrong - I don't think Speedstep voodoo is running. I did a retail install using Chameleon boot loader, and I have the following extensions in /Extra - Disabler, IOACHIBlockStorageInjector, LegacyJMB36xSATA, OpenHaltRestart, and fakesmc

Link to comment
Share on other sites

So memory speed isnt that problem. The MacMini CPU has more Cache so not 100% comparable with yours.

 

I run Geekbench (lastest) 32 Bit with same CPU type E7xxx (maybe a bit higher clocked) at 3.0 GHzand posted

the part of floating point section - main difference MacMini vs your hackintosh.

MacMini > 4100 vs Hackintosh

 

Check every section of that part from my E7xxx with yours - some part (subtest) must get much lower results(Speed) than my, because overall FP for my E7XXX (3.0 GHz) is 5812 (FP, Geekb 32 Bit) !! With 2.66 GHz you should reach at least 4500 FP overall.

 

Can you try to look if all features of the E7xxx are like mine ?

terminal :

sudo -s (asks pwd)

sysctl hw.optional.sse3 =0,1 1=SSE3 available (Geekbench uses beside FPU also SSE/Vector in that FP part)

sysctl machdep.cpu.features

my E7300:

FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM SSE3 MON DSCPL EST TM2 SSSE3 CX16 TPR PDCM SSE4.1

 

So your focus should be:

FP /SSE problems + some background tasks may also be a problem for all benches

Check aktivity when no app is running : cpu % should be near zero ( 98% idle.

 

Attatched my FP part Geekbench 32 Bit E7300 3.0 GHz

Bild_366.jpg

Link to comment
Share on other sites

sysctl hw.optional.sse3 = 1

sysctl machdep.cpu.features = All the ones you listed, but also VMX

 

I also ran a 32bit / 64bit check application that shows as follows:

 

Mac Mini: 64 bit processor, 64 bit EFI, 32 bit kernel boot mode, 32 bit kernel

Hackintosh: 64 bit processor, 64 bit EFI, 32 bit kernel boot mode, 64 bit kernel

 

Is this significant?

 

I'm going to reboot into BIOS and double-check all settings in there.

Link to comment
Share on other sites

Great news!

 

I spent a bit of time looking at the DSDT I used with my build (I used one for an EP45-UD3P) and decided maybe if I try and customise it more towards my own board (an EP45-EXTREME), it might give better results. It did.

 

(note: geekbench-32 results)

 

Old: http://browse.geekbench.ca/geekbench2/view/171966

 

New: http://browse.geekbench.ca/geekbench2/view/174092

 

I'm going to patch it to 10.6.1 and check all is well, but so far this is very pleasing :thumbsup_anim:

 

edit - same result after 10.6.1 installed so looks like the edits to my DSDT fixed the problem! If anyone is interested, I will post up the DSDT I used for future reference as soon as I get back (just going out for a bit)

Link to comment
Share on other sites

I've attached the edited DSDT I've used.

 

I switched back to my main hard disk (Samsung Spinpoint 500gb), which was running 10.6.1 with a 2300~ geekbench score.

 

I removed the graphics / EFI string from the com.apple.boot.plist file, and replaced the DSDT in /Extra with the one from my test system (the one attached to this message) and rebooted.

 

Geekbench then gave me a 4000+ score like before with no other changes needed, so I am very happy :)

 

dsdt.zip

Link to comment
Share on other sites

  • 6 months later...
Hi all,

 

I've done a Lifehacker-guide style install, retail Snow Leopard with the Chameleon boot loader and that's all gone fine.

 

I'm using an E7500 (3.0ghz dual core) and a 9800 GT, and I have a suspicion I still have something to 'fix' with the install before its performing properly. Basically I'm being outperformed by my Mac Mini by a just-noticable amount.

 

The Xbench scores are all over the place, the macmini essentially beats the hackintosh in every area except scores, whereas in actual usage the two machines feel very similar, but the Mini has a slightly quicker feel to the interface and scrolling around in bigger windows.

 

I don't know whether this is simply because I'm running in higher res on the hackintosh, or whether something just isn't optimised as well as the Mini. Right now the mini is at 1680x1050 and the hackintosh at 1920x1200.

 

Xbench and OpenGL results below (1280x960x32) for OpenGL as its the first resolution I can run the same on both computers)

 

Is there anything obviously wrong that I've missed? This is the first hackintosh build I've done and I'm not 100% whether there's much post-install configuration to do.

 

Hi there!

 

I'm newbie in MAC OSX. I'd like to make some questions. There are too much question, but... I know, one a time.

The first one is:

 

*How can I install Snow Leopard on my PC?

 

My Specs:

 

Core 2 Duo E7500 2.93Ghz @ ~4.00Ghz

ASROCK "2Core1333-2.66G" (CPU Bundled Series) - which came with a overclocked cpu @ 2.66, but I changed it.

Gainward GeForce 9500GT 1GB

Supertalent 667Mhz 2Gb + Supertalent 667Mhz 1GB (Dual Channel)

*I don't remember the HDDs*. I don't know if it's so important.

 

As I see, my CPU is Intel. I've read some guides but I didn't achieve anything until now and I'm frustrated.

I've tried lots of pached versions and even a retail Install DVD but I didn't achieve anything.

Only kalyway 10.5.8 (if I remember well) worked for me on my Laptop (Toshiba A300-235) and on VMware 7, but I'm not so satisfied as I would like to.

I tried some times to BOOT one of SL patched version (even the retail with Boot-132) on my PC but without success.

 

 

*Can anyone please help me?

 

Thanks a lot in advance.

Link to comment
Share on other sites

 Share

×
×
  • Create New...