Jump to content
1 post in this topic

Recommended Posts

Hello,

 

This is my first post on the forum.

 

I'm using a low cost hackintosh with the following spec :

- Abit I-45C motherboard

- Pentium E5200 2.5GHz CPU, recently upgraded from 1.8 GHz Celeron M430.

- 2x 1GB DRR5300 Kingston RAMs

 

All is working fine, except that I find the computer rather slow.

 

I used Geekbench on my previous CPU (celeron), and I got around 1770. Looking to the submitted score sounds OK for this type of CPU. Now I upgraded to a new dual-core E5200 CPU, and my geekbench score is .... still 1770 :-(

 

Other benches run on the new CPU :

Cinebench R10 : around 2450 (dual CPU), when all tests on the web say that this CPU reach above 5000 !

Xbench : I currently get a poor overall 105 score, with only 91 for the CPU (

 

Of course I made some search on the forum, and I carrefuly read this topic :

http://www.insanelymac.com/forum/index.php?showtopic=98581

 

But I cannot find any working solutions.

 

Here are my different tests :

 

- SMBIOS output do not report any potential issue with FSB. I tried the smbiosresolver kext to force a fsb setting, but no impact on the benches.

 

- I did not find any valid patched SMBIOS for my motherboard, with hardwired fsb settings, that's why I use the (now deprecated) smbiosresolver.

 

- hacked kernel : I'm using voodoo 9.6 (andy version), with kernel boot flags :

fsb=200 busratio=125 busratiopath=0 -notscverify

with this options, it seems I can correctly set the system fsb, and even overclock the system by passing fsb = 210, 220, 225, etc...

 

BUT, my benches are still SLOW. I certainly miss something, but I do not undertand what.

 

Can anybody help ??

 

- Here is my current xbench result :

 

Results 132.13

System Info

Xbench Version 1.3

System Version 10.5.6 (9G55)

Physical RAM 2048 MB

Model iMac7,1

Drive Type Maxtor 6Y120L0 Maxtor 6Y120L0

CPU Test 91.67

GCD Loop 164.39 8.67 Mops/sec

Floating Point Basic 76.67 1.82 Gflop/sec

vecLib FFT 62.93 2.08 Gflop/sec

Floating Point Library 116.06 20.21 Mops/sec

Thread Test 171.36

Computation 265.85 5.39 Mops/sec, 4 threads

Lock Contention 126.43 5.44 Mlocks/sec, 4 threads

Memory Test 138.02

System 144.75

Allocate 158.92 583.61 Kalloc/sec

Fill 126.21 6136.55 MB/sec

Copy 153.60 3172.65 MB/sec

Stream 131.90

Copy 125.23 2586.53 MB/sec

Scale 124.81 2578.58 MB/sec

Add 141.36 3011.30 MB/sec

Triad 137.85 2948.93 MB/sec

User Interface Test 159.14

Elements 159.14 730.35 refresh/sec

 

 

- Here is my geekbench score (CPU @2812 MHz instead of 2500) :

 

Platform: Mac OS X x86 (32-bit)

Compiler: GCC 4.0.1 (Apple Inc. build 5484)

Operating System: Mac OS X 10.5.6 (Build 9G55)

Model: MacBookPro5,1

Motherboard: http://www.abit.com.tw/ I-45C(Intel i945GC-ICH7) 0.2

Processor: Pentium® Dual-Core CPU E5200 @ 2.50GHz

Processor ID: GenuineIntel Family 6 Model 23 Stepping 6

Logical Processors: 2

Physical Processors: 1

Processor Frequency: 2.50 GHz

L1 Instruction Cache: 32.0 KB

L1 Data Cache: 32.0 KB

L2 Cache: 2.00 MB

L3 Cache: 0.00 B

Bus Frequency: 796 MHz

Memory: 2.00 GB

Memory Type: 0 MHz DDR2 SDRAM

SIMD: 1

BIOS: Phoenix Technologies, LTD 6.00 PG

Processor Model: Pentium E5200

Processor Cores: 2

 

Integer (Score: 1479)

Blowfish single-threaded scalar -- 947, , 41.6 MB/sec

Blowfish multi-threaded scalar -- 1901, , 77.9 MB/sec

Text Compress single-threaded scalar -- 1040, , 3.33 MB/sec

Text Compress multi-threaded scalar -- 1839, , 6.03 MB/sec

Text Decompress single-threaded scalar -- 944, , 3.88 MB/sec

Text Decompress multi-threaded scalar -- 1753, , 6.98 MB/sec

Image Compress single-threaded scalar -- 967, , 8.00 Mpixels/sec

Image Compress multi-threaded scalar -- 1831, , 15.4 Mpixels/sec

Image Decompress single-threaded scalar -- 782, , 13.1 Mpixels/sec

Image Decompress multi-threaded scalar -- 1523, , 24.9 Mpixels/sec

Lua single-threaded scalar -- 1500, , 577.9 Knodes/sec

Lua multi-threaded scalar -- 2726, , 1.05 Mnodes/sec

 

Floating Point (Score: 2634)

Mandelbrot single-threaded scalar -- 1007, , 670.2 Mflops

Mandelbrot multi-threaded scalar -- 1865, , 1.22 Gflops

Dot Product single-threaded scalar -- 1847, , 892.5 Mflops

Dot Product multi-threaded scalar -- 3663, , 1.67 Gflops

Dot Product single-threaded vector -- 1478, , 1.77 Gflops

Dot Product multi-threaded vector -- 3189, , 3.32 Gflops

LU Decomposition single-threaded scalar -- 394, , 350.6 Mflops

LU Decomposition multi-threaded scalar -- 787, , 690.7 Mflops

Primality Test single-threaded scalar -- 1849, , 276.3 Mflops

Primality Test multi-threaded scalar -- 2642, , 490.5 Mflops

Sharpen Image single-threaded scalar -- 2724, , 6.36 Mpixels/sec

Sharpen Image multi-threaded scalar -- 5204, , 12.0 Mpixels/sec

Blur Image single-threaded scalar -- 3505, , 2.77 Mpixels/sec

Blur Image multi-threaded scalar -- 6729, , 5.29 Mpixels/sec

 

Memory (Score: 1930)

Read Sequential single-threaded scalar -- 2211, , 2.71 GB/sec

Write Sequential single-threaded scalar -- 2229, , 1.52 GB/sec

Stdlib Allocate single-threaded scalar -- 1098, , 4.10 Mallocs/sec

Stdlib Write single-threaded scalar -- 2137, , 4.42 GB/sec

Stdlib Copy single-threaded scalar -- 1975, , 2.04 GB/sec

 

Stream (Score: 1571)

Stream Copy single-threaded scalar -- 1591, , 2.18 GB/sec

Stream Copy single-threaded vector -- 1718, , 2.23 GB/sec

Stream Scale single-threaded scalar -- 1762, , 2.29 GB/sec

Stream Scale single-threaded vector -- 1727, , 2.33 GB/sec

Stream Add single-threaded scalar -- 1108, , 1.67 GB/sec

Stream Add single-threaded vector -- 1987, , 2.77 GB/sec

Stream Triad single-threaded scalar -- 1206, , 1.67 GB/sec

Stream Triad single-threaded vector -- 1475, , 2.76 GB/sec

 

- Here is the freq setting from dmesg, showing that kenel options seems to be OK :

 

21/04/09 21:02:16 kernel TSC: Frequency = 2812.500001MHz, FSB frequency = 225.0000MHz, bus ratio = 12.5

 

- Here is the report from sysctl :

 

hw.ncpu: 2

hw.byteorder: 1234

hw.memsize: 2147483648

hw.activecpu: 2

hw.optional.patcher_opts: 0

hw.optional.x86_64: 1

hw.optional.sse4_2: 0

hw.optional.sse4_1: 0

hw.optional.supplementalsse3: 1

hw.optional.sse3: 1

hw.optional.sse2: 1

hw.optional.sse: 1

hw.optional.mmx: 1

hw.optional.floatingpoint: 1

hw.packages: 1

hw.tbfrequency: 1000000000

hw.l2cachesize: 2097152

hw.l1dcachesize: 32768

hw.l1icachesize: 32768

hw.cachelinesize: 64

hw.cpufrequency_max: 2499000000

hw.cpufrequency_min: 2499000000

hw.cpufrequency: 2499000000

hw.busfrequency_max: 796000000

hw.busfrequency_min: 796000000

hw.busfrequency: 796000000

hw.pagesize: 4096

hw.cachesize: 2147483648 0 0 0 0 0 0 0 0 0

hw.cacheconfig: 2 0 0 0 0 0 0 0 0 0

hw.cpufamily: 2028621756

hw.cpu64bit_capable: 1

hw.cpusubtype: 4

hw.cputype: 7

hw.logicalcpu_max: 2

hw.logicalcpu: 2

hw.physicalcpu_max: 1

hw.physicalcpu: 1

hw.cputhreadtype: 1

 

 

I'm available to send other info if needed to understand my issue.

 

Thanks for your help,

 

Romain

 

A new test : if I remove the SMBIOSResolver kext, I got the same bench results, but sysctl reports :

 

hw.cpufrequency_max: 4000000000

hw.cpufrequency_min: 4000000000

hw.cpufrequency: 4000000000

hw.busfrequency_max: 1660000000

hw.busfrequency_min: 1660000000

hw.busfrequency: 1660000000

 

hmm, it seems that SMBIOS tweaks change the way the system see the hardware...

 

Any idea ?

×
×
  • Create New...