VOGONS


Athlon XP performance ratings

Topic actions

First post, by Fish3r

User metadata
Rank Newbie
Rank
Newbie

Hi all, apologies for potentially igniting a 20 year old debate with this question but I've been playing around with early XP (~2003) builds and am a little confused.

I rebuilt my first gaming PC recently (Athlon XP 3000+ Barton, geforce fx 5900xt, 1gb ram) and for fun I got a Pentium 4 rig (2.8ghz Northwood) to see what I was missing out on and it's outright faster than the Athlon machine. My understanding was that a 3000+ should be equivalent to a 3Ghz Pentium 4 but this doesn't seem to be the case. Games are a lot smoother on the P4 machine with both using the same GPU.

I can't seem to find a definitive answer to this online with searches mostly turning up ancient forum arguments. Were the performance ratings referring to the initial slower P4s?

Reply 1 of 24, by johnvosh

User metadata
Rank Member
Rank
Member
Fish3r wrote on 2024-05-04, 22:52:

Hi all, apologies for potentially igniting a 20 year old debate with this question but I've been playing around with early XP (~2003) builds and am a little confused.

I rebuilt my first gaming PC recently (Athlon XP 3000+ Barton, geforce fx 5900xt, 1gb ram) and for fun I got a Pentium 4 rig (2.8ghz Northwood) to see what I was missing out on and it's outright faster than the Athlon machine. My understanding was that a 3000+ should be equivalent to a 3Ghz Pentium 4 but this doesn't seem to be the case. Games are a lot smoother on the P4 machine with both using the same GPU.

I can't seem to find a definitive answer to this online with searches mostly turning up ancient forum arguments. Were the performance ratings referring to the initial slower P4s?

The Athlon XP was marketed using a performance rating (PR) system comparing it to the Thunderbird predecessor core.

"The efficient Athlon XP chips could perform better than similarly-clocked chips from Intel's competing Pentium 4 line-up, which depended on high clock speeds to overcome their low IPC. As a result, AMD believed consumers would be swayed by the megahertz myth. These chips were rated against the Athlon Thunderbird but were popularly compared to the Pentium 4. As a result, the branding became colloquially known as a "Pentium Rating". Maximum PC criticized this as making it more difficult for power users to differentiate between the various Athlon XP chips. For example, two chips could be given the same "PR" branding but have much different engineering (cache size, bus speed, etc), which would affect their performance at different tasks." quoted from Wikipedia (https://en.wikipedia.org/wiki/Performance_Rating)

Reply 2 of 24, by mwdmeyer

User metadata
Rank Oldbie
Rank
Oldbie

Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT.

It also depends on what Pentium 2.8GHz you have. For example if it is the 800MHz FSB with HT I can see that performing better. The Athlon is only clocked at ~2.1GHz.

If it is the 533MHz non-HT version then I would expect them to be more closely matched.

The Barton 2500+ was a great value/cheap chip that could often overclock to 3200+, that is more where the market was at that stage for AMD.

The Athlon 64 was running late.

Vogons Wiki - http://vogonswiki.com

Reply 3 of 24, by Fish3r

User metadata
Rank Newbie
Rank
Newbie
mwdmeyer wrote on 2024-05-05, 00:23:
Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT […]
Show full quote

Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT.

It also depends on what Pentium 2.8GHz you have. For example if it is the 800MHz FSB with HT I can see that performing better. The Athlon is only clocked at ~2.1GHz.

If it is the 533MHz non-HT version then I would expect them to be more closely matched.

The Barton 2500+ was a great value/cheap chip that could often overclock to 3200+, that is more where the market was at that stage for AMD.

The Athlon 64 was running late.

I've got the 800Mhz FSB + HT model, the Athlon machine is performing as I remember but I guess I was expecting less from the p4 due to them being supposedly (literal) hot garbage.

Reply 4 of 24, by Repo Man11

User metadata
Rank Oldbie
Rank
Oldbie

The performance of either brand of CPU could be greatly helped/hindered by the motherboard. Phil had a couple of videos that demonstrated the differences. https://www.youtube.com/watch?v=NECdNX_Wz8k

"I'd rather be rich than stupid" - Jack Handey

Reply 5 of 24, by melbar

User metadata
Rank Oldbie
Rank
Oldbie
Fish3r wrote on 2024-05-05, 01:44:

I've got the 800Mhz FSB + HT model, the Athlon machine is performing as I remember but I guess I was expecting less from the p4 due to them being supposedly (literal) hot garbage.

P4 2.8 HT 800FSB

20240505_09h47m42s_grim.png
Filename
20240505_09h47m42s_grim.png
File size
64.22 KiB
Views
540 views
File license
CC-BY-4.0

#1 K6-2/500, #2 Athlon1200, #3 Celeron1000A, #4 A64-3700, #5 P4HT-3200, #6 P4-2800, #7 Am486DX2-66

Reply 6 of 24, by appiah4

User metadata
Rank l33t++
Rank
l33t++
mwdmeyer wrote on 2024-05-05, 00:23:
Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT […]
Show full quote

Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT.

It also depends on what Pentium 2.8GHz you have. For example if it is the 800MHz FSB with HT I can see that performing better. The Athlon is only clocked at ~2.1GHz.

If it is the 533MHz non-HT version then I would expect them to be more closely matched.

The Barton 2500+ was a great value/cheap chip that could often overclock to 3200+, that is more where the market was at that stage for AMD.

The Athlon 64 was running late.

Athlon64 was running late? Athlon64 absolutely destroyed the entire Intel lineup at the time.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 7 of 24, by stanwebber

User metadata
Rank Member
Rank
Member

no contest between a p4 northwood and an xp barton...the barton will get blown out of the water. i know. i've run the same p4 2.8ghz (and slower 2.6/2.4ghz overclocked) northwood and xp 2500+ barton overclocked to 400fsb and intel clearly runs away with it. the adoption of sse2 vs 3dnow!2 doesn't help amd's case either.

that being said, my main retro rig is still an mp 2800+ barton running on a kt133a board with an isa slot.

Reply 8 of 24, by mwdmeyer

User metadata
Rank Oldbie
Rank
Oldbie
appiah4 wrote on 2024-05-05, 08:48:
mwdmeyer wrote on 2024-05-05, 00:23:
Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT […]
Show full quote

Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT.

It also depends on what Pentium 2.8GHz you have. For example if it is the 800MHz FSB with HT I can see that performing better. The Athlon is only clocked at ~2.1GHz.

If it is the 533MHz non-HT version then I would expect them to be more closely matched.

The Barton 2500+ was a great value/cheap chip that could often overclock to 3200+, that is more where the market was at that stage for AMD.

The Athlon 64 was running late.

Athlon64 was running late? Athlon64 absolutely destroyed the entire Intel lineup at the time.

Yes it was amazing when it came out. But it was late.

Vogons Wiki - http://vogonswiki.com

Reply 9 of 24, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Athlon 64 was released in 2003. Intel had to scramble for 3 years before countering that with Core 2.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 10 of 24, by Shponglefan

User metadata
Rank l33t
Rank
l33t
Fish3r wrote on 2024-05-05, 01:44:

I guess I was expecting less from the p4 due to them being supposedly (literal) hot garbage.

It's unfortunate that the P4 has been tarnished with such a rep, as not all P4 processors are created equal.

Pentium 4 Multi-OS Build
486 DX4-100 with 6 sound cards
486 DX-33 with 5 sound cards

Reply 11 of 24, by Repo Man11

User metadata
Rank Oldbie
Rank
Oldbie

I went with Socket A in August of 2001 by mistake - Tiger Direct sent me a Soyo Socket A motherboard when I ordered a Soyo Super 7 board. After that I stuck with Socket A because it was what I knew, and it was what I could afford. So 2001 to early 2005 when I upgraded to a 939 system. I really didn't bother to know how Intel systems performed relative to what I had as they were completely out of reach price wise.

"I'd rather be rich than stupid" - Jack Handey

Reply 12 of 24, by appiah4

User metadata
Rank l33t++
Rank
l33t++
mwdmeyer wrote on 2024-05-05, 09:15:
appiah4 wrote on 2024-05-05, 08:48:
mwdmeyer wrote on 2024-05-05, 00:23:
Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT […]
Show full quote

Towards the end of their life the Athlon XP did not compete as well as the P4. The P4 got a lot of boosts with FSB, cache and HT.

It also depends on what Pentium 2.8GHz you have. For example if it is the 800MHz FSB with HT I can see that performing better. The Athlon is only clocked at ~2.1GHz.

If it is the 533MHz non-HT version then I would expect them to be more closely matched.

The Barton 2500+ was a great value/cheap chip that could often overclock to 3200+, that is more where the market was at that stage for AMD.

The Athlon 64 was running late.

Athlon64 was running late? Athlon64 absolutely destroyed the entire Intel lineup at the time.

Yes it was amazing when it came out. But it was late.

It was late? Intel took almost 3 years to catch up to Athlon64. I have no idea what some people in this thread are smoking. Am I in an alternate timeline 😁

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 13 of 24, by pixel_workbench

User metadata
Rank Member
Rank
Member

I read a lot of tests and reviews back then, and the Athlon XP ratings were close to what the P4 could do until Intel went to 800mhz FSB and dual channel DDR. After that the P4 was overall faster.

My Videos | Website
P2 400 unlocked / Asus P3B-F / Voodoo3 3k / MX300 + YMF718

Reply 14 of 24, by appiah4

User metadata
Rank l33t++
Rank
l33t++
pixel_workbench wrote on 2024-05-05, 14:08:

I read a lot of tests and reviews back then, and the Athlon XP ratings were close to what the P4 could do until Intel went to 800mhz FSB and dual channel DDR. After that the P4 was overall faster.

That is because those ratings started during the 400/533MHz FSB days and continued as such until Athlon64 got released..

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 15 of 24, by VivienM

User metadata
Rank Oldbie
Rank
Oldbie
pixel_workbench wrote on 2024-05-05, 14:08:

I read a lot of tests and reviews back then, and the Athlon XP ratings were close to what the P4 could do until Intel went to 800mhz FSB and dual channel DDR. After that the P4 was overall faster.

My recollection is that the P4 was in a decent spot in the later Northwood era. They'd finally fixed their chipset challenges and moved past the whole RDRAM situation with the DDR i865/i875 family. And AMD was waiting for K8 to launch...

Then they launched Prescott and everything started going in the wrong direction again, especially in comparison with AMD's 754/939 platforms. The 65nm die shrink was a mild improvement but by that point everybody's attention was turning towards Conroe.

Reply 16 of 24, by mwdmeyer

User metadata
Rank Oldbie
Rank
Oldbie
appiah4 wrote on 2024-05-05, 14:03:
mwdmeyer wrote on 2024-05-05, 09:15:
appiah4 wrote on 2024-05-05, 08:48:

Athlon64 was running late? Athlon64 absolutely destroyed the entire Intel lineup at the time.

Yes it was amazing when it came out. But it was late.

It was late? Intel took almost 3 years to catch up to Athlon64. I have no idea what some people in this thread are smoking. Am I in an alternate timeline 😁

Guys, they were late. AMD expected to release them a lot sooner. I'm not saying that they weren't more powerful than the P4 when they came out.

https://www.zdnet.com/article/athlon-64-delay … ntil-september/

Vogons Wiki - http://vogonswiki.com

Reply 17 of 24, by appiah4

User metadata
Rank l33t++
Rank
l33t++
mwdmeyer wrote on 2024-05-05, 21:39:
appiah4 wrote on 2024-05-05, 14:03:
mwdmeyer wrote on 2024-05-05, 09:15:

Yes it was amazing when it came out. But it was late.

It was late? Intel took almost 3 years to catch up to Athlon64. I have no idea what some people in this thread are smoking. Am I in an alternate timeline 😁

Guys, they were late. AMD expected to release them a lot sooner. I'm not saying that they weren't more powerful than the P4 when they came out.

https://www.zdnet.com/article/athlon-64-delay … ntil-september/

Being delayed is not the same thing as being late. Not only is the delay reported only 1 year and not 3 years as overblown above, but it was also because a) AMD had stock of AthlonXP they wanted to sell to market first b) wanted to focus on Opteron64 with the initial wafers because servers are a more lucrative market. This was a road taken because the AthlonXP was still competitive with P4 and Athlon64 would definitely crush whatever Intel had on their consumer roadmap for 2 more years. Crushing them in the server space before Intel got a foothold with their stupid Itanium 64-bit bullshit was more important. They did not need to rush to the consumer market at all. It was a decision, not a matter of being late. And it was the right call.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 18 of 24, by mwdmeyer

User metadata
Rank Oldbie
Rank
Oldbie
appiah4 wrote on 2024-05-05, 14:03:

Being delayed is not the same thing as being late.

Semantics at this point, what I said was not incorrect.

Vogons Wiki - http://vogonswiki.com

Reply 19 of 24, by bakemono

User metadata
Rank Oldbie
Rank
Oldbie

The performance ratings became more inflated over time. The 1900+ (which I had back in the day) ran at 1600MHz. The 2500+ (which I had later) ran at 1833MHz. How does a mere 233MHz increase translate into 600 marketing points? Increasing cache and bus speeds mitigates memory latency but it doesn't make a whole new CPU.

It's very similar to what happened with Cyrix 6x86 processors. Early on, a 133MHz part was given a PR166 label, then later a 285MHz part was given a far more generous PR400 label.

again another retro game on itch: https://90soft90.itch.io/shmup-salad