SpecInt/SpecFP - Intel vs AMD

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Tonestar, you need to calm way down. Let's debate the actual facts rather than throwing around personal insults. Leave the insults to those without facts.

There are only two types of computer systems in popular use today. They are workstation and server. Workstations are typically used by most people for the occasional Office app, the occasional compilation or rendering (depending on your job), and a great deal of games.

Noone is even discussing the server market or the supercomputer (not popular) market. I never once denied that AMD doesn't have a high end server or supercomputer solution. In fact I supported that statement. But this is entirely irrelevant to the general public, especially those who read through forums such as this. Nearly noone needs a server or supercomputer at home. Sure, there's the occasional scientist or engineer who wants to run simulations or work complex problems, but they are a tiny percentage of those who buy computer systems, and hence those we target for discussions on this board.

Speaking strictly toward the market that matters the most, you don't need an SMP system. It won't help you with the applications and games you use on a daily basis. The best system for the market containing over 99% of people on this earth is the highest performing single processor with the highest performing memory bus.

-Raystonn

-- The center of your digital world --
 
Tonestar? Are you some kind of forum hanger? Have you actually got something intelligent to say? And no I am not FUGGER or Meltdown in disguise. Why did you even join in this debate? HAve you got any REAL facts to offer? Come on whore, show us you've got a brain in that thick skull of yours or go back to your couch infront of the TV and you ball itching stick. Post something useful without acting like an ass and tell us some facts. @sshole.

-----------------

All your RAMBUS are belong t............ ahh screw it
 
What!! What facts did you offer up in your last post? Read my posts, go to the sites. There all facts. I joined this debate because most people are arguing over marginal differences between the 2 manufacturers. Thats because most of them are small time users, who cares which cpu is 7% faster? Most of the tarts here must be small time users like yourself. No one has considered the price/performance combination of a duallie. Duallies are much more affordable (yes Ray there not for gamers ok...for the tenth time) than ever and are a viable solution for semi pro/pro users. So Grizely you [-peep-] wanker is makes alot more sense think about that than marginal differences in single cpu units. You see arselicker thats Intels strong hold. Till AMD provide a reliable solution Intel wins in that ever growing market segment. Not that pussy users like you would ever need one I suppose, if you did have a dually you'd know what is was all about, but then again you could send love letters to your sister on your 486 with no problems...

"Cock-a-doodle-do" is what I say to my girl when I wake her UP in the morning!!
 
>Aren't we both right?

Well, my point was that if you're serious about MP, you'll probably not go with x86. wasn't referring to AMD, because they don't have a solution yet of course.

In theory, there is no difference between theory and practice.
In practice, there is.
 
I read through this entire post, and I conclude...

You're an idiot. You spout off bullsh!t, you throw around personal insults with no provocation, and you have exactly zero allies in this thread. Even the guy who was partly on your side is telling you to put a cork in it. Perhaps it's time you followed his advice.

Kelledin
<font color=red>"Step away from the gimp suit and put your hands on top of your head."</font color=red>
 
Well, you'll just have to take me at my word. I won't be providing my name, company, etc. anytime soon. At least not at your request. My comments stand for themselves. If you dispute the facts, then please address the facts rather then resorting to ad hominem attacks.

You have 12 CNC machines & 2 FMCs? Then you should be familiar with job shop scheduling problems. What sheduling software do you use? And BTW, I could care less how much they're worth. If we want to throw numbers around the projects I've worked on have a few more zeros behind them. But that's as completely irrelaventas the cost of your CNC mills. Frankly, given the level of maturity & judgement you've shown here, I wouldn't trust you to operate a pencil sharpener much less power tools. But that's just me and I digress.

Your finite element analysis is compiled in less then an hour? So, how often do you have to compile your FE software? How long does it take to run? You might look into parallel make files to speed up those compile times on your mighty dual PIII.

Actually, I'm guessing that you mean your FE jobs run in one hour on a single CPU machine. Your not running very big jobs then. How many elements & what kind?

Try some aeroelastic analysis of laminated composite structures. Now there's an FE problem. That's largely what Boeing just built that 96 node athlon cluster for.

Why 36 hours? Because there 24 hours in a day. But todays run depends on how yesterday ended. Boundary condition, ever heard of it? So we need to simulate the previous day. Now we can use some statistical data to initialize the first part of the previous day, but to get realistic conditions for the days end, we need to run the sim for the second half of the first day. The results of that first 12 hours of sim time are junk, except for providing a reasonable starting point for the day of interest. As I said before, it's actually a bit more complicated then that, but you get the idea.

As for the rest of your diatribe...
Yes, dual CPU is barely SMP. It's generally OK for work stations, but its a bottom dweller for much else. And as you keep ignoring, some apps will completely suck on a dual system.

I could care less about the "3d ring". I don't do 3d. Published bechmarks are tremendously useful if ... if ...
well, if you happen to be running the application of the published benchmarks. When I choose a CPU, I benchmark code on it that is important to me. The Intel PIII is so far out of that game it's not even worth discussing further. I haven't had the opportunity to try the P4 yet.

As for facts, I haven't seen a single one from you yet. You're right. I can't back up my numbers & anectdotes for several reasons, that most people with half a brain and a job can probably relate to. I'm relaying my experience. You need to make a judgement call based on that. I highly reccomend that you don't take my word as gospel. But you're a mechanical engineer, so you already know about validating your sources, testing hypothesis, and making informed, objective, unemotional decisions. Right? Good.

You might look the earlier exchange between myself & Raystonn on SSE2 for an an example of how rational people conduct as discussion.

as for your contention that I'm not an engineer. By degree I'm an engineer twice over. You I suspect haven't left grade school yet.

Don't expect another reply from me. I've already wasted to muc time with you.

In theory, there is no difference between theory and practice.
In practice, there is.
 
Cmon ergeorge, your still asking me to believe you use a 16 cpu cluster to figure out your job scheduling issues? My jobs in design (you know, where the brainy people end up) but I can promise anyone that all your formulas just couldn't warrant that sort of computing power. Also go back and read my posts and don't put words into my mouth loser. Nobody here has even visted the sites I mentioned...looks like you all just want to vent your sexual fustrations.I just asked my friend to view this site (he works at MacDoug) as a REAL Aerospace Engineer. I came across laminated composite structures in my first year at uni and so did he. What's the big deal? We both had a good laugh at you. You may have fooled the ignoramousess at this site but any mech or aerospace engineer can pick you for a fraud. If anyone knows egeorge ask him if you can drop in for lunch at work. You'll easily discover he's a fraud. The only one who actually had some constructive intellect was Raystonn but his comprehesion skills were lacking. The rest of you? Pointless insults. Like insecure children with no defence. Egeorge go back and read my posts and comment on the facts published by viatech, Toms Hardware Guide, 3D Labs or Multicad magazine. Don't continue to offer pointless waffle. Your still dismissing the value of P3 in an SMP configuration...you really are a fool. Are all the sites I've put forward wrong? Am I deluded when I activate my stop watch prior to a render and discover it roughly 70-80% faster than my friends comparable single cpu unit? Having a think about what your really saying fool? There will be others here that do understand. And yes some apps don't make use of dual units....but I've said that about 7 times now, so why labour the point? You have provided no URL's, no formulas, no names and no definable structure for your supposed application. You could have saved me alot of effort by just saying "I don't really know".

"Cock-a-doodle-do" is what I say to my girl when I wake her UP in the morning!!
 
Sorry guys, but I have to ask this...

Do you mean your friend works @ McDonnell Douglas!!?
Yes or No will suffice.

In theory, there is no difference between theory and practice.
In practice, there is.
 
ergeorge...what else do people call MacDoug in the aerospace industry. You have obviously been around...NOT!!!Is there any other company that even comes close to the words MacDoug in the avionics industry? Well? Are you sure you don't manufacture womens panties? Are you sure your even employed?

"Cock-a-doodle-do" is what I say to my girl when I wake her UP in the morning!!
 
ergeorge...how can you possibly dismiss the value of SMP with a P3 configuration? Tell me...did you at least go there?(viatech.com) Anyone who visits that site could only come up with 2 conclusions. 1. The entire VIA company are promoting liars....or that your an idiot. If not please do...have a look where it says dual cpu chipsets compared. You may not use design software, raystonn may play games for the most part but you can't deny the huge advantages of a dual P3 unit for CAD/CAM users, animators etc?. Have a look at the 3Dmax sites, MAYA, Softimage, Catia, Ideas, Microstation...the list goes on. And anyone downloading tutorials to program for the SSE 1/2 in the year 2001 should probably get out of programming. I hope you remove that bottle without tearing yourself.

"Cock-a-doodle-do" is what I say to my girl when I wake her UP in the morning!!<P ID="edit"><FONT SIZE=-1><EM>Edited by tonestar on 04/13/01 11:28 PM.</EM></FONT></P>
 
Ok I have established a new policy....I will not be responding to crack whores like Grizely, Kelledin or anyone else that has not put forward some positive technical or finacial concern regarding these issues. You are wasting your time as you have made no viable comments on the issues I have brought forward. Visit the sites or post some of your or simply do nothing. If you would like me to respond you will have to establish some sort of credibillity. If you are not a power user but a pussy user thats ok...as long as you are not a prententious pussy user. Grizely, you have been certified by me as a pretentious pussy user. If you can not establish a worthy use for your computer and prove you have the knowledge base to drive the applications you state I will no longer address you.

"Cock-a-doodle-do" is what I say to my girl when I wake her UP in the morning!!
 
yOU bastard i wasn't even replying to anything about computers how the hell woudl you know? I do not want to get involved. [-peep-] you. I own a business and am more credible than you. you want to face 1 on 1 in real life lets go brother.

-----------------

All your RAMBUS are belong t............ ahh screw it
 
I am afraid this will be the last time I am able to address you Grizely as you still hold a pretentious pussy user title. Please refrain from addressing me until you have established credibility.

"Cock-a-doodle-do" is what I say to my girl when I wake her UP in the morning!!
 
This is quite amusing. A 'newbie' is telling a 'forum patriarch' what to do. This discussion has fallen apart. Everyone needs to get over themselves and drop the egos so we can have a productive discussion on the merits of the only two CPUs this thread is actually about.

-Raystonn

-- The center of your digital world --
 
ROTFLMAO




Still LMAO




Your friend really works for McDonnell Douglas? Really?

Actually we used to call them MacDac back when I worked with those guys. But you see, McDonnell Douglas ceased to exist as a corporate entity back around 1997 when they merged with Boeing. Harry Stonecipher sucked em dry and sold off the corpse.

So, who hasn't been around aviation industry?

In theory, there is no difference between theory and practice.
In practice, there is.
 
You know.. I thought as much, but I started second guessing myself because he sounded so sure of himself. That is quite amusing.

-Raystonn

-- The center of your digital world --
 
Raystonn,
Go back to your original post on Spec. I have to disagree with you on the validity of the comparison. For example, the Athlons were only using PC133 SDRAM in those tests, but I am sure the P4s were using 800 MHz RDRAM. Moreover, there are little info on other components. That is why I would trust results from people like Tom (our host) more. He always makes sure all parties get the best MB and memory, and he always keeps other components (graphic card, hard drive, etc) the same. He also tests with or without SSE2 optimized, which is only fair since most users won't go out to buy all new software. All in all, the numbers on spec.org should not be taken as the last words when we compare processors.
 
Tom's benchmarks aren't very reliable. Take, for example, the SDRAM vs DDR benchmark comparisons: http://www.tomshardware.com/mainboard/00q4/001030/athlon-15.html

The 100MHz memory bus tests using SDRAM and DDR are identical. The speed increase occurs when he looks at the 133MHz bus. Those performance increases are due mostly to an increase in the memory bus, not the DDR memory. Tom states, "The AMD760-platform with Athlon 1200/133 MHz is performing extremely well, scoring almost 11% better than an Athlon 1200/100 on a Asus A7V motherboard." Well gee, you're giving it a 33% memory bus speed increase. SDRAM would give you the same numbers when performing at 133MHz.

Spec is a test of the CPU and the coupled memory subsystem as a whole. Surely if RDRAM were as bad as everyone thinks (hah) it would be hurting the P4's performance, not helping it, right?

The benchmarks most reviewers, including Tom, use for their tests are not optimized for the pipeline of the P4. This has nothign to do even with SSE/SSE2. It has to do with the ordering of instructions. The benchmarks are optimized for tha Athlon, yet not for the P4. I'd be really surprised if a CPU could even manage to keep up under such circumstances, yet the P4 actually wins out in a few of them. That's quite impressive. Spec does not suffer from the imbalanced optimizations rampantly occurring today in benchmarks.

-Raystonn

-- The center of your digital world --
 
It has to do with the ordering of instructions. The benchmarks are optimized for [the] Athlon, yet not for the P4.
Click to expand...
Actually, the benchmarks are probably optimized for the Pentium or Pentium PRO architecture (MMX, SSE, and 3Dnow! ignored). Intel and AMD just happen to optimize their pipelines to closely match the instruction pairing/ordering rules of the baseline i586 or i686, up to the Pentium III and the T-bird respectively.

Surely if RDRAM were as bad as everyone thinks (hah) it would be hurting the P4's performance, not helping it, right?
Click to expand...
We've accepted that RDRAM does well in some situations. Our major beef with it here is its high price (where cheaper DDR might do just as well), and the business practices of Rambus.

Oh well, sorry the thread got trashed. I think we can keep it clean as long as tonestar or the Intel lemmings aren't here...

Kelledin
<font color=red>"Step away from the gimp suit and put your hands on top of your head."</font color=red>
 
No my friend it is you that chooses to supply links that are inaccurate. At this time you owe Tom an apology. His test were right on. What you fail to ignore is at the time Tom did this review there was not a 133 fsb (266 double pumped) SDRAM chipset available for testing. The kt-133a came out after the review you posted a link too. This makes your statement of Tom doing inadequate testing nothing more than a farce. How can he test something that yet did not exist?
I supplied this link in another post to support my claim of ddr ram giving a 6-10% increase, which Tom agreed with in his conclusions. Now you choose to show this as proof of Tom's inaccuracies? How can this be when at the time a motherboard with pc-133 and SDRAM did not yet exist for the 266 FSB athlon's? Further study into this area shows that another review was done after the kt-133a chipset came to market which did support the 266 fsb athlons. And yet still the performance increase's of the better DDR boards was evident.

http://www.tomshardware.com/mainboard/01q1/010314/amd760-08.html#mpeg_4_encoding_flask_mpeg

While all apps do not benefit from the additional bandwith of DDR it is apparent that the more bandwith intensive ones do and Athlon can utilize this increase in memory bandwith unlike the p3. So while an increase in 10% will not be an across the board on all apps, a rounded figure of 6-10% increase in total system performance is a more than fair assesment.

Now is it to say that DDR is the one and only way to go for now? Now this statement I do not fully agree with either it is all cost dependent. When a 6-10% increase is worth the extra 100 bucks give or take then yes it is if not then a SDRAM motherboard will still be a good buy.


A little bit of knowledge is a dangerous thing!
 
Dude, I don't think someone working at MacDonald Douglass would have a 8 year old best friend......

-----------------

All your RAMBUS are belong t............ ahh screw it
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom