AMD or Intel?

  • Thread starter Thread starter Tochiro
  • Start date Start date
John Fryatt said:
I think historically there may have been come compatibility issues between
Intel and other makers in the early days of making 'clone' CPUs. Cyrix
springs to mind.
Click to expand...

I don't think even the old Cyrix chips had a problem with compatibility.
They had smaller cache and slower co-processors (floating point
calculations), so didn't compare well to intel in multimedia tasks and
games, but for windows only work, they were budget and performed well. They
were completely compliant with software - if the processor can handle the
same basic function set (386), then it can run any software designed for
that instruction set. If it didn't comply with the conventions of that
function set, then it wouldn't have even booted windows.
However that has long since been sorted and I don't think you'd have
trouble running anything on an AMD that would run on an Intel CPU.
Click to expand...

Not so much sorted, as evolved - they weren't broken, but cheap budget
alternatives to intel. Cheaper processors are no good if they can't run the
latest games, so didn't sell so well. This forced the companies like AMD and
Cyrix to change their design, or go out of business! The cheaper
alternatives to intel (AMD) now can compete in the same arena and they now
(generally) win!
 
AMD processors and Intel processors both function fine. But AMD has
traditionally been known to give better prices.
Click to expand...

You can't beat a 2.4 GHz Celeron D for $76.

http://www.directron.com/directron/celerond320.html

It is the same as the P4 Prescott chip only it did not pass the rigors
of testing for the P4 specification, eg, L2 pipeline cache size.

In stress tests it compares with my son's 3.2 GHz P4 Prescott. For
example it takes me essentially the same amount of time to compress an
DVD file using DVD Shrink. I was actually surprised how close the
times came.

And I am able to use the Retail Box heatsink - he had to install a
Zalman 7700 to keep his cool.
 
You can't beat a 2.4 GHz Celeron D for $76.

http://www.directron.com/directron/celerond320.html

It is the same as the P4 Prescott chip only it did not pass the rigors
of testing for the P4 specification, eg, L2 pipeline cache size.

In stress tests it compares with my son's 3.2 GHz P4 Prescott. For
example it takes me essentially the same amount of time to compress an
DVD file using DVD Shrink. I was actually surprised how close the
times came.

And I am able to use the Retail Box heatsink - he had to install a
Zalman 7700 to keep his cool.
Click to expand...

But you can beat it. AMD Sempron is higher performance and
lower heat per performance.

DVD shrink is not very well coded at all if it doesnt' run
faster on a 2.4GHz celeron compared to a 3.2GHz P4. I think
there is some other bottleneck in the P4 box. Celeron D is
a good value for an Intel platform, but it IS a "value"
processor that is deliberately reduced in performance
towards that end (for example, they use a lower FSB for only
that reason).
 
DVD shrink is not very well coded at all if it doesnt' run
faster on a 2.4GHz celeron compared to a 3.2GHz P4.
Click to expand...

.... faster on a 3.2GHz P4 compared to a 2.4GHz celeron

Benchmarks all over the 'net will mirror this, while Celeron
performance did improve with higher L2 cache, it's still
clearly a notch below P4 performance.
 
... faster on a 3.2GHz P4 compared to a 2.4GHz celeron

Benchmarks all over the 'net will mirror this, while Celeron
performance did improve with higher L2 cache, it's still
clearly a notch below P4 performance.
Click to expand...

The numbers don't lie.

However, I point out that Shrink for some reason thrashes the disk to
the extent that it is the only app that I have run which heats the HD
up noticeably. Maybe that is the throttle you alluded to earlier.

I am sure his machine is a lot faster than mine on some apps,
especially games - which he plays and I don't play.

Suffice it to say that for a general purpose Internet SOHO machine,
the Celeron D CPU has worked for me, and it only cost $76. If my only
choice had been a P4 for $260 I would have used an AMD chip. But I got
genuine Intel and the mainboard uses the Intel chipset too.

I got Big Brother Inside (TM).
 
kony said:
... faster on a 3.2GHz P4 compared to a 2.4GHz celeron

Benchmarks all over the 'net will mirror this, while Celeron
performance did improve with higher L2 cache, it's still
clearly a notch below P4 performance.
Click to expand...

I have done a real world test with a circuit synthesis tool (CPU intensive)
running for 20 minutes. What I found was the Celeron D 2.4G was 10% slower
than the Pentium 4 2.4G. Using this same test, a P4 2.4G is 20% slower than
1 thread on PD 3.0G. (Of course, you can run two threads on the PD 3G and
both run at full speed, but that is another issue.) So, I would expect the
Celeron D to .9*.8=.72 the performance of a 3G P4 and a bit worse compared
to the 3.2G P4.

Of course, this is just one test that I have done, but it is some
information. I was a bit surprised that the Celeron was a good as it was.

Peter
 
Hi Bob

Yes the Celeron D is a real bargain. And for $76? That's an amazing
price.

If you want a budget PC, AMD proceesors are the way to go. But I forgot
all about those Celerons. They give decent performance at a very nice
price. In fact Celerons were ifrst introduced by Intel to combat AMD
low price point, I believe.


*********************************************************
Sincerely,
Gary Hendricks, Desktop-Video-Guide.com
The best digital video tutorials and how-to guides:
http://www.desktop-video-guide.com
*********************************************************
 
Hi Bob

Yes the Celeron D is a real bargain. And for $76? That's an amazing
price.

If you want a budget PC, AMD proceesors are the way to go. But I forgot
all about those Celerons. They give decent performance at a very nice
price. In fact Celerons were ifrst introduced by Intel to combat AMD
low price point, I believe.

**********************************************************
Sincerely,
Gary Hendricks, Build-Your-Own-Computers.com
The best PC hardware tutorials and how-to guides:
http://www.build-your-own-computers.com
**********************************************************
 
Yes the Celeron D is a real bargain. And for $76? That's an amazing
price.
Click to expand...

More have become available since I bought mine several months ago.

http://www.directron.com/celeron775.html
If you want a budget PC, AMD proceesors are the way to go. But I forgot
all about those Celerons. They give decent performance at a very nice
price. In fact Celerons were ifrst introduced by Intel to combat AMD
low price point, I believe.
Click to expand...

I always thought that Celerons were P4s that could not quite meet the
P4 specs. It is my understanding that my Celeron D is a Prescott chip
but not up to P4 specs so they downsized it and gave it a different
name. I see they are up to 3.06 GHz Celerons for $116, which is half
the cost of a 3.2GHz P4.

When your apps become limited by bottlenecks other than CPU speed,
then the CPU becomes less of an issue. For me the Internet is the most
significant bottleneck, and my Road Runner connection runs downloads
at close to 4000 kbits per second internally. The speed problem comes
from slow websites which no amount of CPU speed can overcome.
 
But I forgot
all about those Celerons. They give decent performance at a very nice
price. In fact Celerons were ifrst introduced by Intel to combat AMD
low price point, I believe.
Click to expand...

Combined with an MCI Intel mainboard with integrated video, LAN,
sound, you can put together a very inexpensive brand-name system.

http://www.directron.com/celerond320.html
http://www.directron.com/msi865gvm2ls.html
http://www.directron.com/kvr40064c3a512.html
http://www.directron.com/097usf.html
http://www.directron.com/wd800jb.html
http://www.directron.com/nd3540a.html

The rest of the parts, eg, KVM, floppy drive, etc are from the system
that was decommissioned.

Add an inexpensive but more than adequate UPS, some high performance
fans, removable drive bays, etc., and you have a real nice system for
a real nice price.

http://www.directron.com/internet750u.html
http://www.directron.com/80l1a.html
http://www.directron.com/kf23.html
 
Bob said:
The numbers don't lie.

However, I point out that Shrink for some reason thrashes the disk to
the extent that it is the only app that I have run which heats the HD
up noticeably. Maybe that is the throttle you alluded to earlier.
Click to expand...

Sounds like you are low on RAM - either less RAM that the other PC, or more
being used by the system and other processes. This indicates that perhaps
you could speed your PC up for other applications too - upgrade your memory!
 
Sounds like you are low on RAM - either less RAM that the other PC, or more
being used by the system and other processes. This indicates that perhaps
you could speed your PC up for other applications too - upgrade your memory!
Click to expand...

I have 512 MB and my son has 1 GB. But I never use more than 185 MB
even with Shrink running. If I have over 300 MB of unused RAM now,
what good is it to add more?
 
Bob said:
I have 512 MB and my son has 1 GB. But I never use more than 185 MB
even with Shrink running. If I have over 300 MB of unused RAM now,
what good is it to add more?
Click to expand...

I sugested upgrading RAM to improve the performance of your own PC. I was
not trying to be critical, but helpful - if whilst encoding DVDs, your PC is
running as slow as a celeron, then perhaps it is always running as slow as a
celeron! I don't think you have over 300MB of unused RAM! If you are using
the windows task manager, it adds up the actual RAM plus Virtual RAM (hard
disk) and gives you a total free ram figure from that!

The reason I say I don't think you have 300MB free RAM is that hard disk
thrashing is a typical sign of excessive virtual memory usage and multimedia
applications typically use a lot of memory. I am not calling you a liar, but
what are you using to measure your RAM usage? I have just booted up my
Windows XP and I have 1 application open - MS Outlook Express, which itself
is using 27MB and I already have a total memory usage of 208MB, so I don't
think your PC is 'really' using only 185MB RAM, I suspect that the software
you have to monitor RAM is probably just reporting 185MB usage! My PC has
1.5GB RAM and when using multimedia applications, it is not uncommon for the
PC to struggle with memory!

My original point was, that you are comparing a Vauxhall Nova 1.4 with a
Vectra V6 and finding they are the same speed! What you haven't told us is
that the Vectra is carrying 4 people and a full boot and driving up hill in
the snow, but the Nova is using aviation fuel and driving down hill!! IE.
Perhaps it is not the processor that is causing the unusual results in your
tests, but other system components dragging the faster Pentium processor
down. I would be very surprised if your RAM usage really is 185MB when
running any multimedia processing application - mine is that high with no
applications running! You could perform a test to eliminate low RAM as a
problem - Open Control Panel -> System. Go to the 'Advanced' tab and click
the 'Settings' button in the 'Performance' section. Choose the advanced tab
and press the 'Change' button at the bottom. For each drive in the dialog
box at the top, choose "No Paging File" and press the Set button (have to
press Set for each drive), then reboot and try to run your multimedia
application again. If the computer runs out of RAM, it will either complain
or just crash! If everything seems OK, then you may find that it runs much
faster.

If that is not the problem, then perhaps it is something simpler - maybe the
application has been setup differently on the 2 PCs, perhaps your PC is
using hard disk for temporary space, where the other PC is using RAM - what
is the RAM usage on the other PC when encoding DVD? If it is much higher,
then it is using the extra RAM on the other PC, but using hard disk
temporary space on yours? Maybe the other PC has a faster hard disk, so is
better at handling the random accesses that the DVD encoder requires.

Whatever the cause, it is not the processor alone that is giving you
spurious results - you are not comparing like with like setups. It could
even be down to the DVD drive itself (or software settings for it) - maybe
it is using temporary hard disk space when reading disks (unlikely, but you
can't rule anything out).
 
I don't think your PC is 'really' using only 185MB RAM, I suspect that the software
you have to monitor RAM is probably just reporting 185MB usage!
Click to expand...

I am using the official Microsoft Windows 2000 Professional Task
Manager. On the Performance page there is a section called Mem Usage,
and it reports 185M.

At the bottom left there is a table with an entry called Commit Charge
Total, and it is the same as the Mem Usage figure.

On the right side there is a table with an entry Physical Memory
Available, and it says around 300MB.
 
Bob said:
I am using the official Microsoft Windows 2000 Professional Task
Manager. On the Performance page there is a section called Mem Usage,
and it reports 185M.

At the bottom left there is a table with an entry called Commit Charge
Total, and it is the same as the Mem Usage figure.

On the right side there is a table with an entry Physical Memory
Available, and it says around 300MB.
Click to expand...

Part of my text that you have cliped said... "If you are using
the windows task manager, it adds up the actual RAM plus Virtual RAM (hard
disk) and gives you a total free memory figure from that!"...

Within the Commit Charge section, there are 3 figures - 'Total' is the
amount of memory (RAM + swap file) that the PC is actually using at any
moment. 'Limit' is the maximum memory (RAM + swap file) available to Windows
and 'Peak' is the highest memory (RAM + swap file) usage level so far. None
of these figures are specifically the amount of physical RAM in your PC.

The reason why you 'appear' to have 300MB of actual RAM available in the
Physical Memory section is because Windows 'swaps out' parts of the
operating system and applications from RAM to the swap file (virtual memory)
on the hard disk. As soon as you start doing anything, windows then 'swaps
in' the cached memory that it requires, which is when you hear the hard disk
thrashing (as you described).

In other words, your free 300MB free is RAM, but that doesn't mean the
operating system only needs 200MB - the 300MB is only free because the
operating system has moved things from the RAM to hard disk temporarily.
More physical RAM would reduces the need for the operating system to use the
Virtual Memory and would speed up your PC!

If you try the test I suggested in my previous post, you will see what I
mean (don't forget to put the virtual memory settings back once you have
seen the difference). With virtual memory turned off, the Commit Charge
section will show quite different numbers - take a note of the limit and
total numbers just after boot-up and after opening a couple of application
with and without virtual memory.
 
I am using the official Microsoft Windows 2000 Professional Task
Manager. On the Performance page there is a section called Mem Usage,
and it reports 185M.

At the bottom left there is a table with an entry called Commit Charge
Total, and it is the same as the Mem Usage figure.

On the right side there is a table with an entry Physical Memory
Available, and it says around 300MB.
Click to expand...

Commit Charge Peak is the relevant entry in Task Manager.
If (hopefully) your ram exceeds that figure, a large part of
the remainder is used for file caching- another very useful
purpose.

There are several ways a bottleneck can be found, like
benching the drives, the memory, and doing a more isolated
test like ripping a disc to one drive while it remains
compressed, the more the better (MPEG2 might be most
relevant to the specific task, but MPEG4 would isolate even
more), then reading in that file, doing a simple re-encode
to MPEG2 onto a 2nd destination hard drive.
 
Commit Charge Peak is the relevant entry in Task Manager.
Click to expand...

What is that figure? Is it the largest RAM usage?
If (hopefully) your ram exceeds that figure, a large part of
the remainder is used for file caching- another very useful
purpose.
Click to expand...

Mine is 325540 MB.

But that is not when Shrink is running.

Are you saying I should note that figure when Shrink is running? If
so, what should I look for?
 
On Fri, 21 Oct 2005 15:01:44 +0100, "GT"

Within the Commit Charge section, there are 3 figures - 'Total' is the
amount of memory (RAM + swap file) that the PC is actually using at any
moment. 'Limit' is the maximum memory (RAM + swap file) available to Windows
Click to expand...

It is the max only within the context of no further pagefile
increase. For example, on a system with 1GB mem, 1GB (fixed
minimum) pagefile and 3GB upper limit on pagefile, that
figure would be around 2GB, rather than 4GB.

and 'Peak' is the highest memory (RAM + swap file) usage level so far. None
of these figures are specifically the amount of physical RAM in your PC.

The reason why you 'appear' to have 300MB of actual RAM available in the
Physical Memory section is because Windows 'swaps out' parts of the
operating system and applications from RAM to the swap file (virtual memory)
on the hard disk. As soon as you start doing anything, windows then 'swaps
in' the cached memory that it requires, which is when you hear the hard disk
thrashing (as you described).
Click to expand...

No, the paged out memory is not discounted from that total.
If he has 300MB free still, it should show up in "Physical
Memory", "Available".

The larger reason why he is only using that amount is that
it's Win2k, not XP. XP could be paired down below that
level of course, but starts out handicapped.

In other words, your free 300MB free is RAM, but that doesn't mean the
operating system only needs 200MB - the 300MB is only free because the
operating system has moved things from the RAM to hard disk temporarily.
More physical RAM would reduces the need for the operating system to use the
Virtual Memory and would speed up your PC!
Click to expand...

No, it really does mean the OS only needs 200MB. WIn2k can
be reduced in background apps and services till it uses less
than 64MB... until you start loading apps. There would
still be minor pagefile activity though, because there will
be a virtual allocation of memory even if it isn't used.
Adding a couple GB of ram won't even stop all pagefile
activity on a box using only 200MB.

If you try the test I suggested in my previous post, you will see what I
mean (don't forget to put the virtual memory settings back once you have
seen the difference). With virtual memory turned off, the Commit Charge
section will show quite different numbers - take a note of the limit and
total numbers just after boot-up and after opening a couple of application
with and without virtual memory.
Click to expand...

IIRC, Win2k doesn't allow turning off virtual memory. Even
so, Commit Charge shouldn't vary except in the "Limit" and
only then because it counts the physical size of the
pagefile as memory, but that does not mean it is using that
memory... is why it's a "limit".

On a WinXP box for example, having freshly booted then ran
two apps (Solitare and Windows Media Player), and misc
things in the background like a few taskbar icon-apps, a
couple of Explorer windows.

512MB RAM and 512MB pagefile,

Commit Charge
Total 138932
Limit 1018392

Then pagefile was turned off, system rebooted and same
things loaded.

Commit Charge
Total 138116
Limit 494108

Physical Memory Available
360512 (figure fluctuated a little, windows likes
doing it's own thing).
 
What is that figure? Is it the largest RAM usage?


Mine is 325540 MB.

But that is not when Shrink is running.

Are you saying I should note that figure when Shrink is running? If
so, what should I look for?
Click to expand...

Compare your Commit Charge, Peak, to the amount of RAM in
the system. If it comes near or exceeds physical ram you
would benefit from more for the *moments* it does exceed...
which could be very often or only rarely, a careful watch of
Task Manager would be needed to better detemine the Commit
Charge Total at any moment (like during DVDShrink(ing)
something).
 
Back
Top