Hrmm, could someone explain to me why Athlon CPUs outperform Intel CPUs (possible gross generalization) in games?
Why is it that Athlon CPUs seem to do better in games?
- Thread starter Antoneo
- Start date
I'd guess due to the Athlon's superior FPU unit, and lower-latency memory access.
Games tend to be branch-intensive also, which naturally favours the 12-stage pipelined A64 over the 20/31 stages of the P4.
Games tend to be branch-intensive also, which naturally favours the 12-stage pipelined A64 over the 20/31 stages of the P4.
It wasn't the case in the early days of the PII/PIII.
When the K7 then hit the scene, the original K7 Athlon performed better in games, simply because it was ~30%-40% faster at the FP math required for the 3D engine, than PII/PIII.
Since then, with the P4s, I'm tempted to argue that it's more due to the P4's performance flaws than any specific property of AMD's Athlons.
Any code exhibiting these flaws were quickly weeded out from all popular benchmarks. Instead, these benchmarks tend to instead give priority to code that runs as well as is even possible, ie max out the P4.
This means Intel's P4s are a bit 'overated' by the general picture most 'popular' benchmarks convey.
The only commonly published benchmarks that break this pattern are game benchmarks. Here a different performance relationship is visable. Thus: "AMD performs better in games".
Actually, that performance advantage is in reality visable in many other applications.
So the short answer would be that the general benchmarking scene is biased towards Intel. Game benchmarking, however, is not.
Netburst only really appears to have an performance advantage in rendering/encoding/multitasking maybe this explains why the AXP 3200+ got such a toasting, possible wasn?t quite worth of the name, but not as bad as it seemed.
I think you need to give credit to AMD for how the K8 has performed even with (if its true what Vee said) fiddle benchmarks.
I think you need to give credit to AMD for how the K8 has performed even with (if its true what Vee said) fiddle benchmarks.
The P4 isn't flawed, it just that Intel insisted on more speed over more work. If they had made the P4 with less/shoter pipelines it would perform a lot more like an Athlon but at the same time the P4 wouldn't be running at 3.4-3.8Ghz
I really think Intel needs to do what AMD did with the first Athlon. They need to throw it all out, the P4 core, using netburst and so on and just start all over.
I think Intel is on the right track.
Just ordered my first Intel after 9 years of AMD only.
Instead of A64 laptop, getting a Pentium M (Not the crappy Pentium 4 Mobile), as gaming reviews put the Dothan 1.8GHz - 2GHz chips at benchmarks comparable to, or faster than, the A64 and P4EE's built today, all at temps that barely require a fan. Couple venders are now making desktop motherboards as well. Easily overclockable. In a short time, with advances in motherboard chipset support, I think Intel will be right there with AMD.
Some links:
http://www.gamepc.com/labs/vie...thandesktop&page=1
http://www.gamepc.com/labs/vie...othangaming&page=1
http://www.gamepc.com/labs/vie...sp?id=dfipm&page=6
i dont have time to read your links right now robs, but ive heard that the dothan can almost keep up with the desktop processors if it is highly overclocked, 2.6GHz. even then, it fails in about half the benchmarks just because of its mobile nature. it cant keep up with desktop processors.
Trust me, I would have said the same thing a week ago.
But, we all are not too old to learn new stuff (I hope).
No problem.
I'll post some quotes and numbers for ya so you can stay here:
From 1st link and 1st tests (Nov 8, 2004):
(nvidia video used)"First off, the Pentium-M chip is a fairly good performer all around. The chip actually puts up numbers on par with Pentium 4 Extreme Edition and Athlon64 FX-55 in gaming, which is no easy feat. In the majority of benchmarks, the Pentium-M at 2.0 GHz can perform roughly on par with a 3.2 GHz P4 or an Athlon64 3200+ processor......(and with regard to low tempsWe didn't need to clock down the CPU, and the CPU performed just as fantastically passively cooled as it does with an active cooling system.."
Doom 3 - Site 3 Timedemo - 1024 x 768
AMD Athlon64 3500+ (2.2 GHz) 137.4 fps
Stock Intel Pentium-M 755 (2.0 GHz) 132.5 fps (lost this one)
Intel Pentium 4 3.4 GHz EE 132.4 fps
Far Cry 1.2 - Volcano Demo - 1024 x 768
Stock Intel Pentium-M 755 (2.0 GHz) 117.7 fps
AMD Athlon64 FX-55 (2.6 GHz) (yes the 2.6GHz chip) 115 fps
Intel Pentium 4 3.4 GHz EE 107.7 fps
From the 2nd link and tests (Nov 11, 2004) because 1st results were unreal and needed confirmed:
(Now with ATi video)
Doom 3 - Site 3 Timedemo - 1024 x 768
Intel Pentium-M 755 (2.0 GHz) 119.9
AMD Athlon64 3500+ (2.2 GHz) 118.6
Intel Pentium 4 550 (3.4 GHz) 118.4
Far Cry 1.3 - Volcano Demo - 1024 x 768
Intel Pentium-M 755 (2.0 GHz) 122.7
Intel Pentium 4 560 (3.6 GHz) 121.9
AMD Athlon64 3500+ (2.2 GHz) 120.3
I could go on, but it's all there in the links.
Another thing AMD has going for it L1 cache 128k of it first gen P4 (all till prescott) had around 16k.
Well, it depends what we're exactly arguing about. So maybe we should take some time to agree on that.
Meanwhile, just a sanity check:
2.0GHz Pentium M 755 costs $435
Desktop mobo for that costs $270
s754 A64 3400+ is $215
nice mobo for that is $70
s393 A64 3500+ is $260
nice mobo for that is $100-$120
P4 3.4 or 550 is $260 and $280
nice mobo for that is $100-$120
3.4/550 P4, 3400+ and 3500+ have at least an edge on the M755 in most benchmarks I've seen. And if you throw some science/engineering phys & math at them, the M755 gets thoroughly trounced.
Both the 3400+ and 3500+ can be the basis for a noiseless PC rather cheaply. The P4 perhaps costs a little bit more to get silent, but hardly anything close to +$300.
Furthermore, the P4 has hyperthreading and the A64s have 64-bit support. M755 has neither.
Of course. But the OP question concerned percieved performance relative to type of software. Not performance relative to clockrate, as you seem to assume.
... - And the P4 is very fickle about what the code look like. I stand my ground on that. And this fact is completely hidden in benchmarks like SysMark, PCMark, Sandra and video benchmarks.
I wasn't arguing value. I was saying that the Pentium M is here to stay, you like it or not.
And yes, it is more expensive now than the others ones. EVERYBODY nows that.
If it is best choise for normal users ? Right now, no. Sometime in the futures it will be a very good candidate.
Just look at the Xtremesystems forum and see how many of the top overclockers are switching from fx55 to the M platform. Can you say 22 seg in Super Pi ?
What gets me is that right now, the P-M is actually cheaper system than AMD when comparing performance for the price as it relates to laptops.
$1500 goes a little farther with P-M high end gaming laptop than it does with a A64+ high end gaming laptop.
Based on desktop price/performance, you would think comparable laptops would give AMD a 20% price advantage. Not so.
How does this relate to gaming in general?
Well when you can now get a laptop with 17" widescreen that pumps out 3dmark 2001 in the 14,000 range, all for under $1500, it has to make you think.
I am now moving video cards around before my Dell 9200 laptop gets here.
It will be the fastest gaming PC compared to the 3 AMD desktops I run (all overclocked AMD XP-M's not A64's).
Because it is portable, it will actually get more gaming use than a desktop, even just around the house.
In 6 months to a year, newer and cheaper desktop/laptop chipsets should appear.
The original question was:
" Hrmm, could someone explain to me why AMD CPUs outperform Intel CPUs (possible gross generalization) in games?"
It mentions nothing about price or value.
Doesn't even mention desktop, so it could have been directed at notebooks.
Hopefully it is now clear to see that Athlons no longer outperform Intel CPU's for gaming.
In fact, 6 months from now the post may read " Hrmm, could someone explain to me why Intel CPUs outperform AMD CPUs (possible gross generalization) in games?
$1500 goes a little farther with P-M high end gaming laptop than it does with a A64+ high end gaming laptop.
Based on desktop price/performance, you would think comparable laptops would give AMD a 20% price advantage. Not so.
How does this relate to gaming in general?
Well when you can now get a laptop with 17" widescreen that pumps out 3dmark 2001 in the 14,000 range, all for under $1500, it has to make you think.
I am now moving video cards around before my Dell 9200 laptop gets here.
It will be the fastest gaming PC compared to the 3 AMD desktops I run (all overclocked AMD XP-M's not A64's).
Because it is portable, it will actually get more gaming use than a desktop, even just around the house.
In 6 months to a year, newer and cheaper desktop/laptop chipsets should appear.
The original question was:
" Hrmm, could someone explain to me why AMD CPUs outperform Intel CPUs (possible gross generalization) in games?"
It mentions nothing about price or value.
Doesn't even mention desktop, so it could have been directed at notebooks.
Hopefully it is now clear to see that Athlons no longer outperform Intel CPU's for gaming.
In fact, 6 months from now the post may read " Hrmm, could someone explain to me why Intel CPUs outperform AMD CPUs (possible gross generalization) in games?
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 384
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter