Is AMD faster than intel

well, if you ask my opinion (which you havent but hey) price will always rule the day, AMD are generally cheaper, probably coz intel think they can get away with higher prices just because they are "intel".

Bang for Buck

my point being, theres no point paying $100 (or £) more if you can oc the chip for $20 extra and a little time

[EDIT] i didnt answer the question, i cant tbh, intel have higher clock speeds, but AMD have a reputation for giving a solution to a problem at a lower cost, with same/close results as intel.

Quote:

---

Originally posted by bisonator
well, if you ask my opinion (which you havent but hey) price will always rule the day, AMD are generally cheaper, probably coz intel think they can get away with higher prices just because they are "intel".

Bang for Buck

my point being, theres no point paying $100 (or £) more if you can oc the chip for $20 extra and a little time

[EDIT] i didnt answer the question, i cant tbh, intel have higher clock speeds, but AMD have a reputation for giving a solution to a problem at a lower cost, with same/close results as intel.

---

Maybe you havent checked prices lately, but AMD's flagship chip is more than Intel's. If your looking purely on an overclockers standpoint, then yes I woould say the edge would go to AMD.

I'll put myself in the line.

Arguing mhz vs mhz is idiotic.
way too simplistic.

Quote:

---

Originally posted by Extol



Maybe you havent checked prices lately, but AMD's flagship chip is more than Intel's. If your looking purely on an overclockers standpoint, then yes I woould say the edge would go to AMD.

---

Of course which eventually they will big time undercut intel like they always do. But the important stuff is the chips that overclock well and are like half the price of a comparitively performing pentium 4.

kirkyg

Intel make their own chipset witch is a +, but the Nforce chipset is looking good so maybe that advantage will be dissapere in the future.

Intel chips have the core covered witch is good (less risk of core crush and better heat transfer), but again new AMD processors should have this too.

The overheatprotection is good, but again VIA and Nforce have this feature too nowdays.

The memory bandwith is the only thing that AMD (VIA, Nforce) can´t match.

Some of the P4 chips are still expensive if you compare to AMD.

Im not going to start to compare CPU:s, but in my opinion for pepole like me that haven´t got that much money to spend on computers. The T-bred 1700+ -> 2400+ are cheap and clock pretty good, so AMD is still atleast my first choice.

Just my 2 cents

//ThAz

Quote:

---

Originally posted by Colossus


There are more then one way to create a fast processor. Intel had chosen the brute force of clockspeed while AMD had taken a different path that doesnt require such a high clockspeed to hit there mark.

---

This is still misleading. The design choices that cause the Athlon to do more work per clock the thee P4 does, are for the most part the same choice that cause the Athlon to not run at the same clock speed as the P4.

You also seem to be implying that designing for high clock rates is somehow inelegant, which couldn't be further from the truth. Historically speaking the processors designed for high clock rates outperform those designed for high "IPC". In fact that was one of the key aspects of the "RISC revolution".

BTW, in the overall scheme of things the Athlon is a *low* IPC, high clock rate processor. For example, it needs to run at around twice the clock rate of Itanium, Power 4, or Alpha processors to achieve similar performance.

Why would the AMD require to run at double to clock rate to even match the speed of the Itanium, Power 4, or Alpha???

If you look at the AMD running at 2.0Ghz (2400+) vs P4 2.0Ghz, the P4 would be slower then the AMD. So how is that possible according to your explanation?

Quote:

---

Originally posted by joerg


As for which is faster clock for clock AMD wins in every case period form 1ghz and up. In PR rating vs P4 clock speed again there is some faulty info floating around here. From 1500+ to 2400+ palamonie and tbred the XP wins and yes i know this is debateable but it is a proven fact that a 1600+ and a 1.6ghz intel the AMD will win. But once AMD changes its PR ratings with the introduction of 333fsb. Past 333 fsb the bartons are way off period so dont even think about them but the 2600+ is abotu the same maybe alittle worse than the p4 2.66 past that it only get worse for AMD but the performance is still only measurabel in fractions of a percent.

---

Not true. On optimized code the P4 performs very similar to an Athlon running at the same clock rate on floating point intensive code. On Integer intensive code, the Athlon generally has the advantage but this is not universally true.

For example, a 2.26 GHz P4 is faster then a 2.25 GHz (2800+) Athlon on 6 of the 14 SPEC FP benchmarks, while the Athlon is faster in the remaining 8. On SPEC Int the Athlon is faster clock for clock on 9 or the 12 benchmarks but the P4 is still faster on 3 so you still cannot claim:

Quote:

---

As for which is faster clock for clock AMD wins in every case period form 1ghz and up.

---

It is generally only on legacy or poorly written code where the Athlon has a big advantage clock for clock. It is certainly easier to write good code for the Athlon, but when the software is properly optimized, the P4 can be quite comparable to the Athlon on a clock for clock basis. (Though still slightly slower IMO)

http://www.spec.org/osg/cpu2000/resu...923-01691.html

http://www.spec.org/osg/cpu2000/resu...827-01589.html

http://www.spec.org/osg/cpu2000/resu...923-01692.html

http://www.spec.org/osg/cpu2000/resu...827-01590.html

I guess it would depend on the program and application..

I just looked at toms and with mp3 encoding the 2.2Ghz beats the 2800+

BUT

With video encoding w/audio the 3.06Ghz beats the 2800+ and the 2400+ is faster then the 2.8Ghz

Quote:

---

Originally posted by Colossus
Why would the AMD require to run at double to clock rate to even match the speed of the Itanium, Power 4, or Alpha???

---

Double the clock rate of the Power 4, Alpha or Itanium.

For example

At 1000 MHz Itanium scores 807 on SPEC Int and 1427 on SPEC FP
At 1250 MHz Alpha scores 845 on SPEC Int and 1019 on SPEC FP
At 1450 MHz Power 4 scores 884 on SPEC Int and 1221 on SPEC FP

(IBM has some better scores, but the consensus is that they used an unrealistic cache configuration. You could set up a Power 4 system the way they did, but it would never happen since you only use 1 out of 8 processors in the system.)

By comparison the Athlon needs to run at 2250 MHz to score 898 on SPEC Int and 782 on SPEC FP

Quote:

---

Originally posted by Colossus
If you look at the AMD running at 2.0Ghz (2400+) vs P4 2.0Ghz, the P4 would be slower then the AMD. So how is that possible according to your explanation? [/B]

---

See my other post on the matter. On properly optimized code the P4 can indeed outperform the Athlon at the same clock rate.

Quote:

---

Originally posted by Colossus
I guess it would depend on the program and application..

---

Very much so.

The Athlon will be faster clock for clock on a majority of applications, and almost all legacy applications, but applications where the P4 *can* be faster are not uncommon.

The only legacy app I can think of where this may hold (P4 similar clock for clock) is quake. This should come as no surprise since John Carmack is know to be one of the best at optimizing his code. Quake III was written properly according to P6 (PPro, PII, PIII) guidelines. That by itself is important, since many in many cases the P4 suffers from the same things that hurt performance eon the P6, but to a much larger degree.

Ok you guys!

Make me a list of apps and benchmarks to run so that I can compare both of my cpus! Try to get benchmarks that are CPU only limited... not having to do with memory bandwidth and fsb, because both machines will not be equally fast... This will be the setup...

Athlon XP @ 2.53ghz
yadda yadda yadda
Geforce 4 TI @ 300 core 650 mem
1024mb Hyper X

VS



Pentium 4 @ 2.53ghz
yadda yadda yadda
Geforce 4 TI @ 300 core 650 mem
1024mb Hyper X


That's probably as close as I can get both computers. Which is why I would like to try out benchmarks that focus solely on CPU clock... since both machines are going to be running diffrent FSB's. 195mhz for the AMD, and 133mhz for the P4.

Ideas ideas ideas!

Quote:

---

Originally posted by lostlilchild6
Ok you guys!

Make me a list of apps and benchmarks to run so that I can compare both of my cpus! Try to get benchmarks that are CPU only limited... not having to do with memory bandwidth and fsb, because both machines will not be equally fast...

---

Almost all real applications depend on memory performance to some degree. Benchmarks that ignore memory performance almost never reflect the real performance of the processor. Using benchmarks like this is one of the largest mistakes that can be made in benchmarking because they do not reflect real applications.

Well i almost gurentee you that the only benchmarks you will find that the AMD will loose are the Spec benchmakrs and some encoding and ecoding benchmakrs simply becasue the software is extremely optimized to take advantage of clockspeed and intel optimization like sse2. Which in my opinion will never be realistic because Software coders simply dont want to optimize for sse2 but benchmarks that do it will see an overwhelming lean in intles favour. Lets say AMD gets somethign like sse2 and then the balance would be shifted back again but i doubt AMD puts the instructions in until software companies start to use them. Also in real world performance the P4 at eqaul clock speeds gets its but kicked by a Athlon in most 95% of situtaions but of course in some things the Intel will always have the lead simply because the software was optomized to run on an Intel.

Heck... Ive always said it... for some odd reason, AMD CPU's seem much quicker than Intel ones :). At least in my opinion. However though... dont get me wrong... a P4 at 3.2ghz is mighty quick indeed :).

Quote:

---

Originally posted by joerg
Well i almost gurentee you that the only benchmarks you will find that the AMD will loose are the Spec benchmakrs and some encoding and ecoding benchmakrs

---

The only form of optimization you can use in SPEC is compiling with a good compiler. Typically that means using Intel's compiler, which produces the fastest code for both Intel and AMD CPUs.

There are relatively few applications that actually need more performance, and these are front and center. There is no reason to expect code that doesn’t need to run any faster to receive any optimization whatsoever.

Quote:

---

Originally posted by joerg

simply becasue the software is extremely optimized to take advantage of clockspeed and intel optimization like sse2.

---

I'm not sure what you are trying to say here. There is not way to "optimize for clock speed". You optimize for the architecture. Clock speed and scaling then governs how fast or slow the code actually runs.

Quote:

---

Originally posted by joerg

Which in my opinion will never be realistic because Software coders simply dont want to optimize for sse2 but benchmarks that do it will see an overwhelming lean in intles favour.

---

As I said, SPEC optimizations are entirely compiler based. All a developer has to do to take advantage of them is use Intel's relatively inexpensive compiler in their development process and pay attention to the data structures they use.

As things stand now, SSE2 makes almost no difference inn performance (5% according to Intel). Remember, all SSE2 is, is MMX and SSE rolled up, with 64-bit operations thrown in. The real key to P4 optimization is paying attention to things like load store forwarding rules, using branch hits, using prefetch, following branch rules, minimizing the use of dependencies on shift instructions, and avoiding consecutive FXCH instructions.

Quote:

---

Originally posted by joerg

Lets say AMD gets somethign like sse2 and then the balance would be shifted back again but i doubt AMD puts the instructions in until software companies start to use them. Also in real world performance the P4 at eqaul clock speeds gets its but kicked by a Athlon in most 95% of situtaions

---

As things stand now, SSE2 makes almost no difference inn performance (5% according to Intel). Remember, all SSE2 is, is MMX and SSE rolled up, with 64-bit operations thrown in. The real key to P4 optimization is paying attention to things like load store forwarding rules, using branch hits, using prefetch, following branch rules, minimizing the use of dependencies on shift instructions, and avoiding consecutive FXCH instructions.

The Athlon may be faster clock for clock 95% of legacy applications that were designed to run properly on a PII 300 but who really cares? On new applications where performance is a concern the P4 is quite competitive clock for clock in most cases. Remember, the SPEC benchmark I cited consisted or 26 separate REAL applications, and the Athlon was most certainly not faster on 95% of them.

Quote:

---

Originally posted by joerg

but of course in some things the Intel will always have the lead simply because the software was optomized to run on an Intel.

---

All but the oldest legacy software is suitably optimized for the Athlon. Are you suggesting it is valid to compare performance based on software that is optimized for the Athlon but no the P4?

Quote:

---

Originally posted by Moridin


Not true. On optimized code the P4 performs very similar to an Athlon running at the same clock rate on floating point intensive code. On Integer intensive code, the Athlon generally has the advantage but this is not universally true.

For example, a 2.26 GHz P4 is faster then a 2.25 GHz (2800+) Athlon on 6 of the 14 SPEC FP benchmarks, while the Athlon is faster in the remaining 8. On SPEC Int the Athlon is faster clock for clock on 9 or the 12 benchmarks but the P4 is still faster on 3 so you still cannot claim:



It is generally only on legacy or poorly written code where the Athlon has a big advantage clock for clock. It is certainly easier to write good code for the Athlon, but when the software is properly optimized, the P4 can be quite comparable to the Athlon on a clock for clock basis. (Though still slightly slower IMO)

http://www.spec.org/osg/cpu2000/resu...923-01691.html

http://www.spec.org/osg/cpu2000/resu...827-01589.html

http://www.spec.org/osg/cpu2000/resu...923-01692.html

http://www.spec.org/osg/cpu2000/resu...827-01590.html

---

Two things i would ask...WHERE THE HELL! did you get those ******** benchmarks? and then why dont you take a look at what benchmarks the athlon wins in. All the important ones.

kirkyg

Quote:

---

Originally posted by kirkyg


Two things i would ask...WHERE THE HELL! did you get those ******** benchmarks? and then why dont you take a look at what benchmarks the athlon wins in. All the important ones.

kirkyg

---

Umm this is the single most important processor benchmark there is. These come straight from the SPEC web site.

SPEC is an organization consisting of all the major chip, system and compiler vendors. Their goal is to develop sound cross platform benchmarks, their members can use to do comparative evaluations of their designs. This benchmark in particular (CPU 2000) is one of the key tools used by processor designers to compare performance across architectures.

SPEC is one of THE best benchmarks out there for gauging the overall capability of a chip. It is used by every major chip company including AMD and Intel. In fact AFAIK the only concrete performance indication AMD has released to date for Hammer have been estimated SPEC scores.

Any real study of processors and processor performance will undoubtedly use SPEC scores at some point, and the ones on this site are the official scores.

Quote:

---

Originally posted by Chronus


Some will most definitely say INTEL has the faster chips, disregarding the huge MHZ abyss between the faster INTEL chip and the faster AMD chip.

INTEL has the faster chip, right now.

It's not a huge lead over AMD, but it's clearly a win for them. On the other hand, if you want a chip that performs almost as well as the 3 GHz INTEL flagship and is comparively cheap, then you'd go for AMD.

Right now, people who have GF4 TI 4800 or ATi Radeon 9700 Pro, a fast HD, fast memories, won't notice any difference between INTEL's 3 GHz and AMD's 3000+. Out of 150 FPS plus, what significance does it have more or less 12 FPS? With both chips, you'll get a terrific performance.

How much are you prepared to pay? If a lot, then go for INTEL.

Chronus

---

Have you compared the price of a XP3000+ Retail at newegg is$579,while a P4 3.06 Retail is $543. So what is this AMD is cheaper stuff about?

Quote:

---

Originally posted by crashed13



Have you compared the price of a XP3000+ Retail at newegg is$579,while a P4 3.06 Retail is $543. So what is this AMD is cheaper stuff about?

---

Thats what I was thinking. :confused:

but that is online recently, in a few weeks/months they will have come down in price. check out all the other products.......

Quote:

---

Originally posted by bisonator
but that is online recently, in a few weeks/months they will have come down in price. check out all the other products.......

---

And so will Intels as they prepare to launch their next chip.

my statement can apply to AMD or INTEL, point being they will all decline to appropriate prices, which as it happens, AMD seems to be cheaper.......1700+ Tbred B?

Quote:

---

Originally posted by bisonator
my statement can apply to AMD or INTEL, point being they will all decline to appropriate prices, which as it happens, AMD seems to be cheaper.......1700+ Tbred B?

---

There all more or less the same price until you go below 2Ghz .

To the person who was kindof dising on the spec benches well first let me say that i am completely with moridin on how great spec is. Spec i by no means another 3dmark or some stupid little benchmark like that is is the benchmark that the big companies like Dell HP Compaq Gateway IBM AMD INTEL and many many other huge corporations pay attention to. The benchmarks are based on instructions that real world applications use suck as 3dsmax and others and they use these instructions to test almost every cpu imaginable. It is one of the few ways to compare cpus suck as power4 sparc itanium P4 athlon hammer and many others.

Now to moridin what do you mean by leagcy application are we talking about unreal2k3 or what because as far as im concerned legacy means programs designed for an OS prior to 98. So i guess most modern games must be leagcy games then becasue the Athlon clock for clock kicks some serious butt. As far as applications go both cpus have strong points and in those strong points one type will usually falt out dominate for instance 3dapplications and video rendering the P4 has from launch been very very strong in thse fields and seems to always take a strong lead there but AMD can also take a strong lead in its own fields.