Slashdot Mirror
Opteron Gaming Benchmarks
bishop writes "Ace's Hardware has published some Unreal Tournament 2003 benchmark results on a 1.8 GHz Opteron 244. The Opteron servers out right now don't have AGP, but this issue has been nullified, literally, through the use of the 'null renderer' option in UT2003 to bypass the display output. At 1.8 GHz, the Opteron manages to outpace all previous Athlon models, though it does still fall behind the 3 GHz Pentium 4 by about 8%." Only 8% slower in performance with a 40% slower clock speed. Not too shabby.
I can't wait till the Athlon-64's come out! Maybe they'll be able to surpass a P4 3ghz?
One Function explains it all ...
bltMegsOfTextureMaps(...) {
return;
}
"hum, nah, that shouldn't mess up the results..."
You know, I find it pretty interesting how vastly AMD's chips outpace the P4 clock-per-clock.It's widely acknowledged that the P4 gets less done per clock than the P3 did. Some people have said that the P4's memory architecture is a disaster, or that it is pipelined TOO deep, but I've got a sort of conspiracy theory about this. Granted I'm not a computer engineer, and know just enough to hang myself with, but here goes:
I think they've manipulated the design so they can deliberately increase the clock rate for marketing reasons, without getting proportionally more performance out. Basically I suppose they've taken the longest paths through the chip and stuck latches all over the place so that the overall cycle time can be reduced, but operations that used to take one clock cycle now may take two. When 1.8 GHz AMDs can nearly match the speed of a 3 GHz P4, I don't think this is an entirely unplausible theory.
On the other hand, maybe the P4 really just is a pig. There was some discussion on usenet a while back about how it takes upward to 2000 clock cycles to enter and exit an interrupt handler on the P4, something which an old 486 could do in ~45 cycle IIRC (and I don't recall exactly, but I think the Athlon today can do an INT/IRET pair in a few hundred cycles). Curious how in hyper-optimizing these chips for the most common cases of execution, performance of these sorts of periphery operations goes all to hell.
That said, I'm really looking forward to having 64-bits on the desktop. =)
I'd pay 50 dollars!
I'll run out and buy one for my blind neighbor.
so I can cream him on tokara forest.
Maybe I can hook up a more advanced version of
that cool braille output doohickey from the
movie sneakers so he can feel the pain with
his fingers.
For every annoying gentoo user, are three even more annoying anti-gentoo crybabies. Take Yosh from #Gimp for example.
I think the Opteron has the potential to do a lot more than be a gaming spectacle, but clearly, between the preliminary SPEC marks, various performance metrics on database machines, coupled with a real bus (hyper transport / alpha - like bus) and being reasonably competitive at a modest clock rate is impressive to say the least. And its ability to chew through legacy un-optimized code is leagues better than Itanium2, which is considering an FX!32 like on the fly "re-compiler" to help with horrific performance.
The novelty here is "I can run junk binaries and be very good at big applications at the same time."
Since when has a potential piece of server iron ever run a gamer benchmark well? The versatility of the Opteron is apparent right out of the floodgate.
This industry needs a bit of boat rocking right now. I will sit back and watch the competition, but this Opteron is quite possibly the first step in de-cheesing and de-crudding the horribly "consumer" oriented PC design.
Sure, this isn't a SPARC,HP-PA, Power4 or a MIPS. But this is so less "gay" than Itanic2. This gives serious power to the masses at a great price. People, lets consider rewarding AMD.
Sun and IBM have both indicated that they will be coming up with designs around the Opteron recently. I remember wanting that Power4 a while back. I remember hoping Apple de-vaporized the 970. Now the Opteron is here, and well, I don't want to wait for a Power4 or a 970, lets take this baby and run with it.
You know, it's always saddened me that there aren't more multi-cpu motherboards for general consumption.
There's plenty of us that would stick another £200 CPU in the box every now and then as the cash came in.
I mean, I've already got more memory than I ever use and too much HD storage, I want CPUs I never use either.
I like my DUAL 1.2 Ghz p3 mb. I put it together for £300 or so and it's done sterling service, surpassing my normal 18 month upgrade timetable and not looking like it's about to be retired any time soon. (until DoomIII & HL2 I guess 8).
I'll be looking to buy a nice 4-8 way Opteron system and populate it as I go along.
I think Intel missed an opportunity crippling their processors to be non SMP. I hope there's an engineering reason why. I can't really think of a good reason why otherwise, surely they don't make *that much* on gambling that I'll buy a expensive Pentium if I can't have a 4 way Celeron set-up.
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
Oddly enough, Anandtech's numbers show the complete opposite: The Opteron beat ALL other CPUs at ALL games. Including the P4 3.0C in UT2K3. Take a look:
http://anandtech.com/cpu/showdoc.html?i=1818&p =6
It's interesting to see what's almost a disclaimer after the benchmark results saying that it is not a final benchmark result basically because it wasn't using the full capabilities of the chip.
There has been no benchmarks so far except perhaps for a few Linux benchmarks where the Opterion has been nearly as optimised by the software as the Pentium, Athalon or Xeon simply because the software producers and developers haven't caught up to the chip, and it's a big transfer as well.
In my opinion, we won't see any true representative benchmarks for months yet, perhaps when the A64 has been released.
Only 8% slower in performance with a 40% slower clock speed. Not too shabby.
but what is the price difference?
Don't forget that this is 32-bit code. When the code gets ported to 64 bits, performance will be even better. The benefits come from AMD64 having lots of additional registers over x86-32.
Pharmacist:Oh, that drug only treats cancer, but this one cures it.
Retarded idiot: How much is the drug that cures it? Because my money is good and I would rather remain on the treatment for a lifetime for a perceived value than actually fix the problem.
The fat greasy poor sexless live-at-parent's-home geeks without real jobs, money or responsibilities and their shallow pockets are the reason the PC is such a piece of shit right now. You cheap assed donkeys would rather reward cheap ass companies like VIA and ABIT with cheap junk motherboards, cheap ass components and you show your friends your flaming crud rig can bench ok with 3Dmark. No, you wouldn't want to reward a Silicon Valley company such as AMD for engineering excellence. Keep throwing your short supply of dollars to the people how make half-assed shit and the half-assed shit makers Win! Hooray! Mobocracy by the venue of pseudo-intelligent mediocritomatons buying lame chink shit all the time.
Your bottom lines are : I can't afford anything and I care about gamer benchmarks.
I agree with you completely. That's why I use a Power Mac.
And this explains the SPEC marks kicking the P4's ass how?
Dell Precision WorkStation 650 (3.06 GHz Xeon)
SPECINTpeak 1138
SPECFP peak 1063
Opteron 144
SPECINTpeak 1170
SPECFP peak 1219
How is this any indication of anything other than total and complete supremacy over the P4?
Also, when considering I/O, from GiGE cards and storage subsystems, something gamer kiddies rarely do, the waiting 2000 cycles for an interrupt to get properly handled is bizarre. This is a very bad formula for servicing I/Os efficiently.
With hyper transport, also, the 1-4 way performance will cream the Xeon MP. Only the 2-way Xeon and the 2-way Opteron have any chance of being in real performance contention, until you want 16GB of system memory or to run a 64 bit optimized application.
Also, the relocation of most of the important parts of the chipset onto the CPU is another added bonus.
People, if you want to cling to your smoke and mirrors and Intel mythology, go ahead. I want to see a benchmark where: a Mylex RAID card is at 50% stress level from strenuous md-I/O, a large compile job is being performed, such as make world or making the linux kernel, the box is passing about 150,000 packets per second through it - in one NIC, out the other - all at once. When I get my hands on one of these things I'll let you know, but here is a hint: The Opteron system is not only going to beat the Xeon box, but probably mop the floor with it.
The most bizarre thing is that Intel submits to SPEC and they make the compiler the use!! AMD doesn't even have its own compiler yet, and they are clearly defeating Intel Xeons. Also note that the Power 4 and SPARC are being defeated by the Opteron at this point. And certainly you aren't suggesting that Power 4 and SPARC systems are shitty are you? Also note that while the Itanium 2 can beat the Opteron, it can't run junk binaries or IA32 code for shit.
Intel is betting that retards and gamers will continue to drive their sales. But seeing as Sun and IBM have both announced systems based on Opteron, maybe this time the masses wont force more garbage down our throats.
BTW, I still have a P3-1400Mhz 512K cache chip at home. It was cheap, and for whatever reason, most probably the P4 sucking so bad, I have had zero desire to get a new system this past year. The Opteron is the first glimmer of hope in some time.
All this ass kicking is doled out on the first spin of a chip with a clock rate 40% less than of the hackneyed, boring cheap high profit margin for Intel Pentium 4. Keep helping Intel have enough cash lying around to make Blue Man Group commercials to further brainwash retards. That's where your money goes. To Intel, to pay for completely shitty commercials because it makes you feel good. Just don't look at SPEC marks, because the truth hurts.
The second reason is because of the interrupts. In a dual-CPU machine, one CPU can be getting hammered by interrupts, and you still have another to run other code, such as your GUI.
True to a point. There must be a working/efficient APIC and the OS has to properly manipulate this for both CPUs to service interrupts. While I strongly agree with a 2-way system being leagues better than "faster" single boxes, there are big-time problems with various OSes including Linux where say, putting a GigE card in, then putting it in promiscuous mode, then sending 200,000 packets per second at the box causing complete and total livelock. The problem comes with the ability for interrupts to very easily monopolize the whole system. Other architectures have solved this problem with things like crossbars, more efficient busses, etc.
I am hoping the Opteron will handle IO better - a lot better. We have a very nice Serverworks dual P3 implementation right now with 64bit 66MHz PCI cards (E1000, 3C996). It is fairly easy to send enough traffic to livelock the box. The CPU is cold to the touch, its not doing anything, but it is live locked. So, with the APIC turned off, 1 CPU is 100% used up by interrupts. With it on, 2 CPUs are 100% used up by interrupts. Interrupts steering isn't an availability guarantee is my point.
I have solved problems with a polling feature available in FreeBSD, where you can set the maximum amount of CPU time (interrupts fall under "system") that the network driver can take up. The point is to have the ability to set a fraction of the CPU aside for userland. This is very useful for network applications that require seeing lots of traffic, like Snort, but need CPU to do other things.
The point is that the PC architecture is seriously flawed in many ways. If you crack open the hood of a Juniper you'll see a "PC brain" running FreeBSD, but they solved these crazy architectural flaws by having network processors. Everything is co-processed in Juniper land.
Unfortunately, even the best NICs with interrupts coalescing and all the CPU cycle saving junk turned on doesn't help for network applications that need to see it all.
I'm hoping for better IO and better 2/4 way performance from Opteron, and I think it will deliver. Given that UT2003 is completely un-optimized for Opteron yet Opteron can play ball with what it considers "junk binaries" is impressive.
All in all, I must say, the minimum number of CPUs for anyone who is in the slightest technically inclined is 2.
I'm just going to ignore the rest of your rant - I didn't even read it.
I know, its okay, your attention span is too short and low-IQ troglodytes such as yourself have to pace their learning.
I don't fucking care if you don't want to be edified. What a fucking retard. You care enough to write back and say some shit, then say, "but the rest of your rant, being mostly based in fact and in experience, is invalid to me."
The fact you announce you don't read comments is like a 2 year old little rotten spoiled brat fucking kid that sticks his hands over his ears and screams, nah nah nah nah nah nah nah.
What a fucking loser. You are fat, sexless and live in your parent's basement, I know, but you can take out your frustrations on your plushies, no here.
Oh yeah, and why the AC?
I have a karma capped account for other things.
I wasn't upset... The more we go back and forth, the more of a testament to the fact you are a fucking know-nothing bitch "T.B." loser I can build on the pages of Slashcrap.
How long is the unemployment line these days? I'd figure I'd ask someone who knew. And I know you mom gets pissed when you don't go to bed by 2AM, and the Affexor probably keeps you up, but its her house, her rules.
You sexually dysfunctional lunatic loser freak.
BTW, fuckhead, even as an AC I'm rarely ignored. Your inability to be technical in anyway is a formula for being ignored. Suck on deeze nuts and fo shizzle, sad fat sexless bitch.
I don't know what to make of Piers Haken's overgeneralizations. On the one hand, Piers Haken's argument is invalid. But on the other hand, one of the most widespread manifestations of the craziness of our world is racialism. One of my objectives is to embrace the cause of self-determination and recognize the leading role and clearer understanding of those people for whom the quintessential struggle is an encompassing liberation movement against the totality of extremism.
I have the following to say to the assertion that the Queen of England heads up the international drug cartel: Baloney! At first, he just wanted to dispense outright misinformation and flashlight-under-the-chin ghost stories. Then, he tried to convince loquacious self-promoters that there is absolutely nothing they can do to better their lot in life besides joining him. Who knows what he'll do next? The complete answer to that question is a long, sad story. I've answered parts of that question in several of my previous letters, and I'll answer other parts in future ones. For now, I'll just say that his operatives believe that there is an international Communist conspiracy to sap and impurify all of our precious bodily fluids. I say to them, "Prove it" -- not that they'll be able to, of course, but because Piers Haken thinks that he can ignore rules, laws, and protocol without repercussion. However, he simply regurgitates the empty arguments that have been fed to him over the years. His methods are much subtler now than ever before. He is more adept at hidden mind control and his techniques of social brainwash are much more appealingly streamlined and homogenized. Piers Haken's agendas are popular among insensitive schmucks, but that doesn't mean the rest of us have to accept them.
There's something fishy about Piers Haken's prevarications. I think he's up to something, something vulgar and perhaps even inarticulate. Plainly stated, Piers Haken's cat's-paws have learned their scripts well, and the rhetoric comes gushing forth with little provocation. Piers Haken has called innocent children annoying utopians to their faces. This was not a momentary aberration or a slip of the tongue, and hence, we can safely say that he respects nothing and no one. Piers Haken's followers probably don't realize that, because it's not mentioned in the funny papers or in the movies. Nevertheless, he says that arriving at a true state of comprehension is too difficult and/or time-consuming. What he means by this, of course, is that he wants free reign to deprive people of dignity and autonomy.
I discussed this topic in a previous letter, so I will not go into great detail now, but we mustn't let him infringe upon our most important constitutional rights. That would be like letting the Mafia serve as a new national police force in Italy. To state it in a more sophisticated manner, Piers Haken doesn't use words for communication or for exchanging information. He uses them to disarm, to hypnotize, to mislead, and to deceive.
If I have a bias, it is only against disgusting hoodlums who obstruct important things. Mutual efforts against subversive feudalism are not just an educational process designed to teach people that most of Piers Haken's principles are slanted in the same ideological fashion, with large amounts of emotional exaggeration and general ignorance. These efforts also serve as a beacon, warning the world of the obtrusive consequences of Piers Haken's demented ideas. In contrast, if I withheld my feelings on this matter, I'd be no less savage than Piers Haken.
Let me back up a little: The space remaining in this letter will not suffice even to enumerate the ways in which he has tried to convince innocent children to follow a path that leads only to a life of crime, disappointment, and destruction. We should agree on definitions before saying anything further about Piers Haken's gruesome, impetuous announcements. For starters, let's say that "insurrectionism" is "that which makes Piers Haken yearn to stretch credulity beyond the breaking point." He is
Whatever freak. It is more fun to be free from the chains of festering self aprecating self congratulating mediocritomaton mobocratic group thinking bleating Slashdotting Slashbot fucking Windows Loving zealot fuck-heads like you. I'd rather use acerbic wit than be a conformist. Anonymity frees me form the mass-murdering pathetic sociopaths like you who are angry at the world because your lack of charisma makes gay or straight sex hard to come by. You strike me as the type to use saran wrap on your keyboard, you know, so the key-plungers will depress because spooge tends to make them stick.
Your continued patronization to this thread further illustrates my point and my grievances against you. My line of thinking is still correct and yours is still wrong propagating from the root of this half dead tree (with me being the live part).
Fat. Sexless. Different. Slashbotting Anime Loving Windows 'Hacker.'
Yellow is the color of piss, Spoogeman.
Vae Victus - and that would be you, spoogeman.
McPee Pants in the hizz-zouse. Where you come up wid dat shizz noz?
You're hillarious. Please, continue...
You are under the illusion that I am somehow at your beck and call. My previous treatises on your lack of technical insight and the decimation of your feeble character have been more than adequate to show even the most dullard plebian that you are unfit to be espousing about much of anything.
I don't see that you've said much of anything, especially concerning the original point - CPU instruction pipeline lengths. In fact, most of what you've said has been a reflection of your character. Maybe that's why you've chosen to remain anonymous? Surely your karma rating isn't that important to you?
I didn't have to say much of anything after my original point. No use being right more than once. But I'm getting paid right now, so I'll have a bit of fun.
About the length of the pipeline comment you made, it simply reveals how little you know about CPUs. 40% less deep pipeline and 40% less clock doesn't mean the performance is going to similar. While that may appear to be the case with the Opteron and the P4-3Ghz-HT, it isn't.
With the SPEC peak ratings, the submitter is allowed to use compiler tweaks to achieve a near perfect situation for the CPU given the code it has to process. Why is it that it then the SPEC-peak for the P4 3GHz and the Opteron 1.8GHz far off? Opteron is only seriously challenged by the Itanium2. The point is that while branch prediction may be accurate, it can't save you on a really deep pipeline often enough to justify the depth. And you walk off with your "fast" processor that needs tweaked binaries to run well.
Try doing a make world on FreeBSD, as another commenter has suggested, and even throw in some absurd optimizations to gcc 3.2.2 (use FreeBSD 5 for this) (for athlon-xp: -march=athlon-xp -m3dnow -msse -mfpmath=sse -mmmx -O3 -pipe -fforce-addr -fomit-frame-pointer -funroll-loops -frerun-cse-after-loop -frerun-loop-opt -falign-functions=4 -maccumulate-outgoing-args -ffast-math -fprefetch-loop-arrays ; for pentium4: -march=pentium4 -O3 -pipe -fomit-frame-pointer -fforce-addr -falign-functions=4 -fprefetch-loop-arrays)
You could even do a make -j 6. Even with a faster drive, the Pentium 4 will lose this. There was another comment in this thread confirming this. I have confirmed this. If anything the deeper pipeline is much worse peer level of depth added, the law of diminishing returns not only kicks in bit for every increase in clock per added depth you get even less per unit of depth added, unless the goal is to fake being fast to get people to buy "3 Ghz" chips.
Not to mention I/O. Now you have to interrupts the CPU to get important things done and now you have more instructions in flight. Why do you think Intel's super duper mega chip is not nearly as deep? Well, people who buy real iron look at SPEC rates and application performance before and probably don't pay any attention to the clock at all.
So Windows kiddie, any more bits of incomplete and incorrect wisdom you think you might have to share?
The situation with the gigE cards isn't entirely the fault of the OS (although the OS certainly isn't blameless). GigE cards simply generate *swamps* of interrupts. However, the nicer gigE cards give you the ability to turn on interrupt coallescing, where the card will queue up several packets before it sends an interrupt.
That does, of course, increase the latency, but unless you're in a clustering environment, you can generally handle a small hit in latency much better than you can handle a swarm of interrupts.
On the note of embedded PC architecture, yesterday I helped my boss crack open an old GPS to try and replace the interal battery. What did I see staring back at me? A 25 Mhz i386! It was, of course, in a *much* smaller package than the "regular" 386's were in, I suppose that Intel makes (or made) them on a smaller process for a while for projects like that. And I recall $15,000 Livingston PM3's which used an AMD 486. : )
steve
Oh, you're not stuck, you're just unable to let go of the onion rings.
The interrupt coalescing doesn't help much. I have verified this personally. FreeBSD's device polling does, where the card must ask for interrupts through a polling routine.
Also, I know of several implementations in hardware (Juniper, 6509, Extreme, etc) where full blast ingress on all ports isn't a problem - offloading the problem to network coprocessors isn't a luxury anymore, it's a necessity.
Hopefully the Opteron implementation will deal with interrupts more gracefully than garbage PCs. Im not impressed with non-APIC I/O, where CPU 0 is the only interrupt handler, or APIC, where n- number of CPUs can handle interrupts, but the interrupt steering does nothing to alleviate livelocking the OS. The people who write network drivers should all consider, will there be ANY user CPU time leftover after I'm done servicing interupts. They don't care apparently.