AMD 64 X2 4200 dual core seems slow

  • Thread starter Thread starter John Perry
  • Start date Start date
J

John Perry

I have run Norton's Passmark Performance Test 5 from Systemworks 2005
and have some strange results after upgrading my Athlon Venice to a
dual core with 2GB DDR 400 RAM: This is CPU Benchmark

Athlon 3000 (2GHz) Venice: 373.9
Athlon 4200 (2.2Ghz) Dual Core 423.9
Lenovo T60 2400 Dual Core 816.3

Why of why does my ThinkPad with just 512MB RAM appear to be twice as
fast on the overall CPU test?

I'd appreciate comments
 
I have run Norton's Passmark Performance Test 5 from Systemworks 2005
and have some strange results after upgrading my Athlon Venice to a
dual core with 2GB DDR 400 RAM: This is CPU Benchmark

Athlon 3000 (2GHz) Venice: 373.9
Athlon 4200 (2.2Ghz) Dual Core 423.9
Lenovo T60 2400 Dual Core 816.3

Why of why does my ThinkPad with just 512MB RAM appear to be twice as
fast on the overall CPU test?

I'd appreciate comments
Click to expand...

The A64 has Cool & Quiet which drops the clock speed when the processor is
running a light load. It sounds like your system isn't raising the clock
to full speed when it's running that benchmark. In Windows there is a
power management control panel, you should check to see if its set to some
low power mode (I'm a Linux user so I can't tell you what the proper
choice is in Windows but it should be obvious). It's also possible that
you are missing the drivers that control Coll & Quiet or that your BIOS
needs an update.
 
The A64 has Cool & Quiet which drops the clock speed when the processor is
running a light load. It sounds like your system isn't raising the clock
to full speed when it's running that benchmark. In Windows there is a
power management control panel, you should check to see if its set to some
low power mode (I'm a Linux user so I can't tell you what the proper
choice is in Windows but it should be obvious). It's also possible that
you are missing the drivers that control Coll & Quiet or that your BIOS
needs an update.
Click to expand...

Thanks for that. I downloaded the latest drivers for the CPU, and
then the processor ran really hot and I had a blue screen.

Next I installed the AMD power monitor, and set the processor into
laptop mode, and now it runs cool at 33C not 45C! And then I ran
performance test again and the result was 410.

Now, what is interesting is that performance test only used I core and
the frequency went up from 1000 to 2200 and the voltage from 1.1 to
1.35V on that core, but the other core had nothing to do.

Conclusion is that I have the 10% improvement going from 2GHz to
2.2GHz, but I still have a spare core that is not being used, so the
performance is really around 800.

Now, how the Lenovo reports 816, is beyond me.

Thanks again for the tips!
 
John said:
Thanks for that. I downloaded the latest drivers for the CPU, and
then the processor ran really hot and I had a blue screen.

Next I installed the AMD power monitor, and set the processor into
laptop mode, and now it runs cool at 33C not 45C! And then I ran
performance test again and the result was 410.

Now, what is interesting is that performance test only used I core and
the frequency went up from 1000 to 2200 and the voltage from 1.1 to
1.35V on that core, but the other core had nothing to do.

Conclusion is that I have the 10% improvement going from 2GHz to
2.2GHz, but I still have a spare core that is not being used, so the
performance is really around 800.

Now, how the Lenovo reports 816, is beyond me.

Thanks again for the tips!
Click to expand...

Note, the test program may be single threaded. The OS will then dump
the various sundry windows processes onto the other core as
it likes but likely those processes still won't work that one
out much compared to the load prog. which is made by design to use lots of
cpu. Try running two instantiations of the test program at the same time.
Mark
 
At full speed, you should not go over about 50'C and you shouldn't get a
BSOD. Something is wrong with the setup or the cooling.

Normal idle temps for an x2 should be around 30-35'C in a room with
ambient temperature of 20'C and proper case & CPU cooling.

Perhaps your heatsink is not properly seated?

Or your case has very inadequate fan exhaust and you have a big hotspot
inside the case?

A friend once had a heat problem with his system. It turned out the
power supply fan was thermal controlled and would run at very slow
speeds if the power supply wasn't being taxed.

The ambient case temp was over 40'C so the heatsink could not cool the
CPU enough under load. He didn't have a case fan, just the single power
supply fan, so the case had poor air flow. There were mounting holes for
a case fan, so I just added a $15 80mm fan to exhaust the hot air and
all was well again.

But that's running at half speed again, so you're losing performance and
not using the second core. Double whammy!
Note, the test program may be single threaded. The OS will then dump
the various sundry windows processes onto the other core as
it likes but likely those processes still won't work that one
out much compared to the load prog. which is made by design to use lots of
cpu.
Click to expand...

That's not completely accurate.

Even with a single threaded app, Windows XP will distribute the process
across the two cores, and it's rarely a 50/50 split. What actually
happens is the thread is distributed as the process priority is
executed, so the load will vary on each core as the program runs and
time slices with the other processes that are running at the same time.

Sometimes when you upgrade a processor, XP doesn't upgrade it's kernel
hardware access layer (HAL) to work with the extra core - it doesn't
"see" your dual-core CPU. There are two kernels, one is the single
processor (SP) and the other is the multiple processor (MP) HAL.

Also, you need to ensure your BIOS supports dual-core processors, and
that if you're running XP Home Edition, you have service pack 2 (SP2)
installed.

Now to check, bring up the Task Manager. Look under the Performance tab
and if it only shows one graph for your CPU then XP is not running the
multiple processor HAL kernel and it's not using the second core.

You can also check in Device Manager where it should be showing 2
processors. Expand the Computer options and you should see "APCI
multiprocessor PC" for the MP HAL.
If it says "Advanced Configuration and Power Interface(ACPI) PC", then
you're still running the SP HAL.

So if it looks like you're using just one core, you need to open the
Device Manager and uninstall the CPU from the system, and restart your
computer. That should force XP to update the HAL and you should have two
CPUs shown in Device Manager after the restart. You may have to remove
the CPU running in Safe Mode...I can't remember if it'll work from the
normal desktop mode.
 
Look at the true speed of both desktop CPUs, 4200 only runs 10% faster.
The new 4200 uses added features for better performance (like 3DNOW
professional?)
But that does not mean that test programs will take full advantage of newer
instructions.
Your older 3000 comes with 1024K level two cache, the 4200 only has 512K per
core.
What CPU is in the ThinkPad?
Is it an AMD dual core running at 2.4GHz?
 
At full speed, you should not go over about 50'C and you shouldn't get a
BSOD. Something is wrong with the setup or the cooling.

Normal idle temps for an x2 should be around 30-35'C in a room with
ambient temperature of 20'C and proper case & CPU cooling.
Click to expand...

I second the temperature range you quote, same processor. The hottest I
ever saw when loading the processor with things to do was 52DegC. Normal
web-browsing and the such is around 33DecC. If you enable the fans to spin
lower for a bit quieter time (especially the CPU one) then the temperatures
do go up (I think it's called QFan or something).
 
Conclusion is that I have the 10% improvement going from 2GHz to
2.2GHz, but I still have a spare core that is not being used, so the
performance is really around 800.
Click to expand...

Most benchmarks only use 1 CPU.

A dual core VS a single core at same MHz is about 90% faster on
multi-threaded apps or when multi-tasking several apps.

Ed
 
Thanks for comments:

Look at the true speed of both desktop CPUs, 4200 only runs 10% faster.
The new 4200 uses added features for better performance (like 3DNOW
professional?)
But that does not mean that test programs will take full advantage of newer
instructions.
Your older 3000 comes with 1024K level two cache, the 4200 only has 512K per
core.
What CPU is in the ThinkPad?
Click to expand...

Intel T2400 Core Duo 1.83GHz
Is it an AMD dual core running at 2.4GHz?
Click to expand...

1.35GHz in each core

Also the test only used one core in the AMD - I do not know if the
Intel processor was clever enough to use both.
 
Back
Top