my new computer is slower than my 6.5 years old computer

Do you have a question? Post it now! No Registration Necessary.  Now with pictures!

Threaded View

hi everyone

i have put together a new computer to replace my old computer that died
after 6.5 years of service. my new computer incredibly is slower than my
old computer. i've already tweaked the bios. like newton being hit on the
head by a falling apple, i've been scraching my head to look for an

i've cheated a bit by overclocking my old computer. it was too slow and
even if the overclocking kills it, i won't be losing much. i don't
overclock new computers because they should have enough power.

i've further cheated by replacing the original cpu with the fastest the
mobo could support. when the ide-sata converter died, i just swapped the
10,000 rpm raptor hard disk with a new 7200 rpm ide hard drive in the new

but the old motherboard is an abit kr7 that could not even run usb 2.0.
it could not support agp 8x, so i had to run an ati 800 on 4x. the mobo
was so old that i feel it could not possibly be faster than my new
computer regardless of how i have tweaked it.

my new computer is based on an abit ax78 with phenom quad core 9550. it
runs on ddr2 1066, compared with ddr 266. it has dual channel memory,
compared with single channel. it has pcie, compared with agp 4x. it has
geforce 8500 gt, compared with ati 800.

apparently i was wrong when i thought that quad core 9550 runs like a cpu
of 9550 mhz and should be three times faster than my old cpu. i found out
that intel core 2 e6600 is faster than my cpu even though it has only two
cores and should be equivalent to 6600 mhz.

the strange thing is when i surf the internet with a browser, i'm 100
percent certain that my new computer is faster than my old computer. but
my new computer could not even reach 500 when benchmarked by passmark. my
old computer scored 515.

unless there's something wrong with passmark rating 6.1, we can expect
amd to declare bankrupcy anytime. my quad core 9550 computer was outrun
by an intel computer with only one core.

anyway, i've made up my mind. my next computer will have an intel cpu.


Re: my new computer is slower than my 6.5 years old computer

On Sat, 13 Sep 2008 01:49:04 +1000, sharonpst thoughfully wrote:

Quoted text here. Click to load it

I'm not familiar with Passmark did you compare your results to other PCs
or does to other baseline results.  

Phenoms run a a slow cpu speed, ie 2.4 vs. 3.6, if Passmark returns a 500
based on cpu speed alone it would be lower.  However, if it compares your
PC to other PC platforms you should see improvement in application speed.

Re: my new computer is slower than my 6.5 years old computer

jaster wrote:

Quoted text here. Click to load it

Is Passmark multi-core aware or is it benchmarking just one core?  The
webpage says "Supports hyper-threading, and multiple CPUs testing".....
is that multiple physical CPUs or multiple core CPUs?

Re: my new computer is slower than my 6.5 years old computer

sharonpst wrote:
Quoted text here. Click to load it

It is too bad you didn't give any details about your old processor,
as I could have attempted to describe the performance difference to
expect between the old processor and the new one.

First of all, the "model number" of a processor now, is meaningless.
To AMD, the "9550" number is just a number. The AMD numbers that
had the "6400+" with the plus sign on the end, were P.R. or
Performance Rating numbers. AMD has stopped doing that, since
their competitor is no longer a P4, but some other (Core2)

The best way to rate them now, is to use benchmarks, and there
are many to be found around the net.

I would try a couple things.

1) Disable Cool N' Quiet. That should cause the processor to
    run at full speed all the time. Then I'd try a copy of
    CPUZ from, to verify the clock rates involved.
    The core should be at 2200MHz. Make sure the control-alt-delete
    (Task Manager) shows four performance graphs for the four

    When only a single core is working for you, a 2200MHz core
    should be equal to a 3.3GHz P4, in terms of benchmarks.

    If I use SuperPI and calculate 1 million digits of PI, it
    takes about 50 seconds. You should be able to beat my
    performance with seconds to spare. SuperPI only uses a
    single core. I'm running a 3.1GHz P4. Give it a try.

2) Usually, there is a bit of "driver nonsense" to do for
    AMD processors. This thread gives some examples. You
    could probably turn Cool N' Quiet back on before attending
    to this.

Note - I cannot get to respond right now, and
so I had to use an archived copy. It is better to read the
current version of the thread, to get the latest advice.

3) You can look up video card parameters here. /

    The X800 I picked, has a 256 bit wide memory array, while
    the 8500 GT uses a 128 bit wide memory array. I think given
    you were using the X800 before, you might have picked a
    slightly higher end video card for the new machine. The
    8500 GT might be what I'd pick if I didn't plan on gaming.

The 8500 GT isn't at the bottom, but it's pretty close.,741.html


Re: my new computer is slower than my 6.5 years old computer

Quoted text here. Click to load it

I dont know the answers here, but can note a couple of points. First a
quick google shows passmark to be a tester that runs under windows, so
win itself will have a big impact on performance. I know from
experience that different versions of win can have a huge impact on
performance, with generally newer versions being very much slower.

Other point is that a vid card can make a huge difference for some
apps. A P3 machine tested with 2 cards:
basic ISA card -> 4fps
AGP card -> 91 fps

Your situation might be due to something else entirely of course, I
dont know.


Re: my new computer is slower than my 6.5 years old computer

hi paul, nt

both of you are right.

before my old computer died, i benchmarked it with passmark. when it ran
xp, it scored 380. when it ran win2k, it scored 515. that's the reason
why i wanted to hang onto my win2k. it runs faster and gives fewer

i did what paul suggested and removed cool n quiet. the score notched up
from 497 to 502. i installed an ahci for the sata drives. it went up to

the problem lies with win2k sp4. it only sees two cores for some reasons.
so i have been benchmarking only two cores. i have tried to benchmark in
winxp to see whether it could see four cores in task manager. but i have
trouble running winxp because of driver problems. i'm only running 32bit
windows, never mind about 64bit o/s.

i remember reading somewhere that amd's quad cores are really dual cores.
that also might be the trouble.

i've spent so much time and energy that i believe i have reached the
point of no return. i'm really tempted to go back to a second-hand 32bit
mobo where everything is guaranteed to work.

but that would only be delaying the inevitable. when the hardware and
software become too old, unfortunately we have to go through the painful
process of upgrading. lots of time spent booting and rebooting.

for the record, my old cpu was a mobile athlon. i used an adaptor to
force it to run at top speed all the time. it was the equivalent of 3200+
but i overclocked it to 3500+. the old graphic card was not an x800, but
a slower ati 800 bought two years ago. with that ancient set up, my old
computer managed to score 515.

thanks everyone for their contributions wrote in

Quoted text here. Click to load it

Re: my new computer is slower than my 6.5 years old computer

sharonpst wrote:
Quoted text here. Click to load it

AFAIK, Win2K is limited to two cores, so the other two will likely not
show up. WinXP will show all four. (Only one socket is involved
in this question, so either WinXP Home or WinXP Pro would work and
show all four cores of the quad.) Since I use Win2K, this is one
reason a quad may not be an immediate priority for me. The OS upgrade
is a significant expense.

The Phenom is a true quad core, with the four cores sitting on
the same silicon die, with a large L3 shared by them. The
Intel Core2 Quad, on the other hand, uses two silicon dies,
and the FSBs are connected together inside the processor package.
This is one reason, when testing scaling with a good scaling
application, an Intel Quad scales to 3.5x (chokes), while the Phenom
scales to 4x (perfect scaling). But since Intel Quads are more powerful
when the clock rate and IPC are considered, the lack of perfect
scaling doesn't really matter. Intel wins by brute force.

As for Passmark, I like simple benchmarks, to separate the
various aspects of the system. SuperPI is a good way to do
a quick test of CPU and memory performance. You can tweak,
and immediately see how much of an impact it makes to the
SuperPI compute time. SuperPI uses only one core, and for
me at least, is more indicative of what to expect in normal
desktop usage. I don't have a lot of multicore aware software.
(Photoshop might be the only one.) SuperPI should not be
testing your disk, for example.

I'm on a AthlonXP-M right now (my backup machine) at 3200+, and
SuperPI is 45 seconds for 1 million digits. My 3.1GHz
P4 does it in 45 to 50 seconds. The 50 second mark, is
with my AV software present.

(The xtremesystems download seems unavailable, so I used this.)


Re: my new computer is slower than my 6.5 years old computer

hi paul

please don't tell me that i have made another discovery. who in their
right mind would buy a quad core cpu if they were told in advance that
they would need to buy new software that are multicore aware.

i was told that multicore is the way to go because manufacturers cannot
keep pushing up the mhz limit because of the extreme heat generated. how
they increase the efficiency of the cpu's is their problems.

the understanding has always been that this would be transparent to
users. everything will be downward compatible and users can keep their
old software. right?

thanks again

Quoted text here. Click to load it

Re: my new computer is slower than my 6.5 years old computer

sharonpst wrote:
Quoted text here. Click to load it

There is a difference, between "compatible" and "efficient".

Compatible means, all the old software runs. Even if you were
on Win2K and only two of the cores worked, your software still

Efficiency, on the other hand, assumes that every part of the
system is working to use the hardware to best advantage. That
might not be the case here. Only certain kinds of software
will get the best benefit from the quad. Or, only certain
usage patterns will extract the maximum good from the new

Before giving up on the processor, I suspect you may not
have tried all the adjustment knobs on your system. Not
every aspect of a system, comes optimally set up when you
get it.

For example, you could visit this thread and see about
optimizations for the CPU driver. There are 650 posts
to the thread, which means the situation may have
changed since the advice was first written. With SP3
out for WinXP, things could have changed a bit. I don't
know if the AMD driver page has had any good advice added
to it or not (it hasn't had good advice in the past).

(Phenom driver downloads),,30_182_871_15259,00.html

A minor thing, is there could be an option in the BIOS,
to operate the RAM "ganged" or "unganged". As I understand
it, Phenom has two independent 64 bit memory interfaces.
For quad core operation with lots of different tasks running,
"unganged" operation is probably best and may be the default.
Each memory channel works independently, so two tasks can read
or write memory at the same time (subject to the addresses
being to the different channels). An alternative, is to use
"ganged" mode, which might make more bandwidth available
to a single task. The difference might improve a single
task's performance by a few percentage points.

The Abit AX78 manual (not available from the Abit site!),
takes a pass on addressing the memory subsystem on the board,
so there is no info on whether "DCT" (a DRAM Controller),
"ganged", "unganged" are mentioned in a BIOS screen. I
downloaded an Asus manual, and it offers
"DCT Unganged Mode" [Auto or Always], and that
kind of terminology doesn't do anything for my
blood pressure :-) Lots of waffling on the subject. It
would seem they only want to support the Unganged option.

So there are still a few things to play with.

I think you're right, in that a 3500+ rated single core
AthlonXP, would be very close to a single core on the
Phenom at 2200MHz. Scaling that Phenom core clock speed based
on a rough guess at IPC, that is about 3300+ or so. So
the trick now, is to get more than one core to work for
you. If only one core is busy, the Phenom is only 25%
used in a sense.

If you had a 6400+ X2, that runs at 3.2GHz, faster than
the 2.2Ghz of a 9550 core. 3.2*1.5 = 4.8GHz P4 or so.
Just by way of comparison. When all cores are used to
the max on those two processors though, the Phenom
gives 4*2200=8800 "units", while the 6400+ gives
2*3200=6400 "units". So for the marketeers, the Phenom
is "faster", while for ordinary users, the 6400+
is snappier for a lot of stuff.


Quoted text here. Click to load it

Re: my new computer is slower than my 6.5 years old computer

hi paul

thanks much for your advice.

every weekend, i will be tweaking my systems. i'm sure they would be
running very efficiently after a few years. by that time, the mobo might
die again, and i will be back to square one. but that's life.

thanks again

Quoted text here. Click to load it

Re: my new computer is slower than my 6.5 years old computer

Motherboard: MSI P45 Neo3-FR MS-7514
Memory: DDR2-800
Disk: ST3500823AS
Video: Gigabyte GV-NX86T256H

After CPU test, PassMark Rating (Partial) is 461.1
After 2D Graphics test, PassMark Rating (Partial) is 545.9
After 3D Graphics test, PassMark Rating (Partial) is 581.8
After Memory test, PassMark Rating (Partial) is 718.6
After Disk test, PassMark Rating (Partial) is 819.6
After CD Standard test, PassMark Rating is 842.3

PassMark(TM) PerformanceTest 6.1 ( )
Results generated on:  Saturday, September 13, 2008

Benchmark Results

Test Name: This Computer
CPU - Integer Math: 237.9
CPU - Floating Point Math: 925.7
CPU - Find Prime Numbers: 907.8
CPU - SSE/3DNow!: 5953.5
CPU - Compression: 6269.8
CPU - Encryption: 34.5
CPU - Image Rotation: 1182.0
CPU - String Sorting: 3936.1
Graphics 2D - Lines: 174.3
Graphics 2D - Rectangles: 79.8
Graphics 2D - Shapes: 6.6
Graphics 2D - Fonts and Text: 321.2
Graphics 2D - GUI: 387.3
Graphics 3D - Simple: 686.5
Graphics 3D - Medium: 185.6
Graphics 3D - Complex: 61.2
Memory - Allocate Small Block: 3159.4
Memory - Read Cached: 2252.4
Memory - Read Uncached: 2240.8
Memory - Write: 1993.7
Memory - Large RAM: 595.8
Disk - Sequential Read: 78.0
Disk - Sequential Write: 64.6
Disk - Random Seek + RW: 3.0
CD - Read: 2.1
CPU Mark: 1690.1
2D Graphics Mark: 608.3
Memory Mark: 757.2
Disk Mark: 526.9
CD Mark: 259.0
3D Graphics Mark: 298.7
PassMark Rating: 842.3

System information: This Computer
CPU Manufacturer: GenuineIntel
Number of CPU: 1
 (2 Core(s)/CPU, 1 Logical(s)/Core)
CPU Type: Intel(R) Core(TM)2 Duo CPU     E8400  @ 3.00GHz
CPU Speed: 3007.6 MHz
Cache size: 6144KB
O/S: Windows 2000
Total RAM: 2047.1 MB.
Available RAM: 1137.4 MB.
Video settings: 1280x1024x32
Video driver:
 BIOS: Version 60.84.5E.00.00
 DATE: 07/09/07
Drive Letter: E
Total Disk Space: 7.8 GBytes
Cluster Size: 4.0 KBytes
File system: NTFS


Quoted text here. Click to load it

Site Timeline