Not at all.
I based this on in house testing of real world apps that we use for HTPC's, HDTV and gaming. My results mean nothing though to most, so I tried to find others with similar results. Hard to find comparisons models later.
Not at all.
I based this on in house testing of real world apps that we use for HTPC's, HDTV and gaming. My results mean nothing though to most, so I tried to find others with similar results. Hard to find comparisons models later.
Mean nothing because maybe you are not testing properly. Everybody rather to trust more in bechmarks of sites like anandtech, xbitlabs (the show dscription and details of the systems, the benchmarks, the drivers, how they test,etc,etc) and others than in some homemade testing. The bechmarks says otherwise. In gaming as showed the Athlon XP is a huge bottleneck , maybe it's the videocard is bottlenecking the opteron in gaming that's why you don't see the difference. I can tell you from my personal tests I REALLY SEE THE DIFFERENCE and it's just HUGE (an they are pretty much in concordance with all the test I have seen in good reputation sites). It's going to be very if not impossible to find someone with your similar results.
BINGO!!!
Unless you also invest in a high end video card, which you have, (and most that are debating AXP Vs Sempron would not be doing, since they are budget minded), you will not see much performance increase in gaming going from Socket A to Sempron...
Other real world results for HDTV are seen at AVSforums. People are struggling with new high end systems trying to play WMV-HD smoothly, yet with a properly setup AXP, it will play smoothly. Properly setup 939 will of course be better, but in HDTV area, it is only about 5% better.
Will you email be faster or web browsing be faster? Not that anyone would notice.
I still see no other first hand benches.
As far as Athlon XP (Barton not T-Bred) versus cored Sempron (Paris not T-Bred). Don't trade in the XP-M Barton just yet....not if you have the 35 or 45 watt Mobiles. Theres plenty of room to OC those particular CPU.
Is the socket 939 Sempron combo faster than the 754 combo? Yup! Is the 64 bit Sempron faster than the 32 bit XP-Barton...yup!!
It seems so tempting, but I still can't (cost effectiveness-wise) justify upgrading for the near future after (so far) seeing such little visible (still way over 30 FPS) difference in gaming. Sure you say that a minute or so more time spent for encoding is worth 120 bucks....I personally don't see it.
As far as Athlon XP (Barton not T-Bred) versus cored Sempron (Paris not T-Bred). Don't trade in the XP-M Barton just yet....not if you have the 35 or 45 watt Mobiles. Theres plenty of room to OC those particular CPU.
Is the socket 939 Sempron combo faster than the 754 combo? Yup! Is the 64 bit Sempron faster than the 32 bit XP-Barton...yup!!
It seems so tempting, but I still can't (cost effectiveness-wise) justify upgrading for the near future after (so far) seeing such little visible (still way over 30 FPS) difference in gaming. Sure you say that a minute or so more time spent for encoding is worth 120 bucks....I personally don't see it.
I am not agree.
The Athlon XP is a huge bottleneck for 6800GS/GT or X800XL/XT and up videocards, Those are not considered hihgh end anymore but mainstream. So there is no point in your reply.
THG tests are made with 6800GT videocard it shows tha barton @2GHz bottlenecking this card in a bad way, you can see that because in games like DOOM3 and UT2004 the Semprom 2800+ is (running 400MHz slower) beats it by nearly 30%. That's huge bottlenecking.
Besides I am not talking only about gaming. Look at the other benchies. The athlon XP gets whiped by AMD64 CPUs.
I am obviously not talking about those apps, because those are not CPU intensive (and are most internet-conection dependant), if you run only those apps even a celeron will be OK, so there is no point in this comment.
Faster is not the term (faster would be 1-5% diff), It's much faster (15-30% diff) clock for clock.
More benchies:
http://techreport.com/reviews/2004q1/athlonxp-m-2500/index.x?pg=12
http://techreport.com/reviews/2003q3/workstation/index.x?pg=11
"However, the Opteron itself showed quite a bit of promise, tying or beating its older sibling, the Athlon XP 3200+, "
As maluckey pointed out, but I neglected too, basing my results on comparing several 's maximum overclocked AXP-M's to several maximum overclocked Opteron 939's.
No one is saying socket A keeps up.
Even though above link shows it blowing away (not my term) A64 in some tests, we all know that on the whole even 754 is faster. But, not much is the point. In a game that is getting 30fps, 10% is only 3 FPS, and even 20% is only 6 FPS, which most would say is NOT worth nothing, and noting close to WOW, it blows it away. Burning a DVD in 10 minutes or 12 minutes also is not blowing away anything.
http://techreport.com/reviews/2004q1/athlonxp-m-2500/index.x?pg=12
http://techreport.com/reviews/2003q3/workstation/index.x?pg=11
"However, the Opteron itself showed quite a bit of promise, tying or beating its older sibling, the Athlon XP 3200+, "
As maluckey pointed out, but I neglected too, basing my results on comparing several 's maximum overclocked AXP-M's to several maximum overclocked Opteron 939's.
No one is saying socket A keeps up.
Even though above link shows it blowing away (not my term) A64 in some tests, we all know that on the whole even 754 is faster. But, not much is the point. In a game that is getting 30fps, 10% is only 3 FPS, and even 20% is only 6 FPS, which most would say is NOT worth nothing, and noting close to WOW, it blows it away. Burning a DVD in 10 minutes or 12 minutes also is not blowing away anything.
At same clockspeeds? I see you don't read well. I am saying the AMD64 blow away AXP CPus CLOCK for CLOCK, now what happen if we OC the A64 at the same clockspeed the barton is OCED it is going to get whipped again by 30%+.
Burning a DVD depends on the burning speed not on CPU speed, it is not a CPU intensive app (look at the CPU utilization when you burn, it should be low), the burners use DMA remember?.
Posted way back that am basing this on Opteron running at max overclockable speeds of 2.7GHz and 3Ghz, so it is an apples to apples comparison when using stock HSF's and running systems at the best they are easily capable of running.
AXP-M running 2100MHz-2200MHz so should I slow down the Opteron's from 3Ghz to 2100MHz?
I give up. Sorry to drift off in this thread. Just pointing out that some owners of AXP that have moved to next gen were taking in by the hype, and results are no where close. I'll leave it at that. You can find many more benchmark comparisons by googling AXP-M and socket 754 Vs socket A.
Zap
Elite Member
- Oct 13, 1999
- 22,377
- 7
- 81
Doesn't matter. At the provided link the Opteron with DDR266 still beat the AXP with DDR400. Here are some numbers for the memory test.
AXP dual channel DDR400 2587
Opt single channel DDR266 2764
Opt single channel DDR333 3186
Opt dual channel DDR266 4019
I know these are synthetic benchmarks, but the Opteron got more out of slow memory than the AXP can get out of the best memory. How about an Opteron with dual channel DDR400? We can only guess at the numbers (7000?).
carlosd, be nice.
Regardless of the direction this thread took I think you have your answer. Disregarding cost, availability, future upgradability, platforms...The socket 754 Sempron will overall be measureably faster than the socket A Sempron at the same true MHz and even with ½ the L2 cache. Also the Palermo Semprons can usually overclock enough to meet or beat any socket A overclock.
Can we all agree on the above statements?
So you are saying that diference of an opteron at 3GHz is low when compared to an AXP @2200, that is the most ridiculous thing I have ever read. Again How come the Semprom 2800+ (1.6GHz) is beating the BArton @2GHz running by nearly 30%, those are framerate differences of 26 and 50 FPSs respectively. An opteron @3GHz (given it`s not limited by video card) will be a world faster than the AXP, I can't imagine how much but can give an Ideal looking at the FX-57 results in DOOM3 and UT2004 and comparing it to a AXP barton 3200+ (@2.2GHz), the difference in performance is huge.
http://www23.tomshardware.com/cpu.html?modelx=33&model1=238&model2=269&chart=69
Originally posted by: Zap
Doesn't matter. At the provided link the Opteron with DDR266 still beat the AXP with DDR400. Here are some numbers for the memory test.
AXP dual channel DDR400 2587
Opt single channel DDR266 2764
Opt single channel DDR333 3186
Opt dual channel DDR266 4019
I know these are synthetic benchmarks, but the Opteron got more out of slow memory than the AXP can get out of the best memory. How about an Opteron with dual channel DDR400? We can only guess at the numbers (7000?).
carlosd, be nice.
The socket 754 Sempron will overall be measureably faster than the socket A Sempron at the same true MHz and even with ½ the L2 cache. Also the Palermo Semprons can usually overclock enough to meet or beat any socket A overclock.
Can we all agree on the above statements?
I agree TOTALLY with all of your statements.
So you are saying that diference of an opteron at 3GHz is low when compared to an AXP @2200, that is the most ridiculous thing I have ever read. Again How come the Semprom 2800+ (1.6GHz) is beating the BArton @2GHz running by nearly 30%, those are framerate differences of 26 and 50 FPSs respectively. An opteron @3GHz (given it`s not limited by video card) will be a world faster than the AXP, I can't imagine how much but can give an Ideal looking at the FX-57 results in DOOM3 and UT2004 and comparing it to a AXP barton 3200+ (@2.2GHz), the difference in performance is huge.
http://www23.tomshardware.com/cpu.html?modelx=33&model1=238&model2=269&chart=69
[/quote]
You can slow down the opteron at those clock speeds and still will be an abuse for the against AXP. It will still get whipped by the opteron. At least as I said before your opteron is limited by your videocard or you are running a light-CPU game (there are not much arount lately , all the of motherfuckers are CPU HOGS!).
PingSpike
Lifer
- Feb 25, 2004
- 21,765
- 614
- 126
Meh. My point was more that once you sold your old equipment, which still fetches a pretty good price you can essentially make this relatively mild upgrade free of charge. The change isn't totally amazing in some benchmarks, but in others it is signifigant.
But if you're happy with the XP setup, sticking with it is a perfectly valid position. It certainly involves a lot less work!
Yes. Exactly!
The difference you link to is the extreme.
Clock for Clock 2200 Vs 2200 and less CPU intense games, like FarCry or Half-Life 2, and things change.
Even in that Doom III chart example at 2200MHz, will most notice a difference in a game that runs 80 FPS compared to one that runs 100 FPS? I think not. Blown away by the speed difference? HaHa. Most in a blind test would not be able to tell which was running 100 FPS. That is real world results.
OK, those are DOOM3 and UT2004 games , which are not so new (neither are HL2 or FC), in more demmanding games the will be even more bottlenecking and I am thinking in the upcoming games which will be even more CPU demanding. (Also those are only Average FPS), If they could show the FPSs continously every second you will see what is the drop in framerates, there you will notice the difference. YES in average the framerates in the FX57 is 110FPS but ot will never fall down under 60. I the AXP config you can see some slowdowns because of the CPU bottlenecking , the framerate is in average 80FPS but it can fall abruply as low as 60 or under 30 FPSs in more demanding games because that CPU can not handle the data stream as efficienly as the AMD64 CPUs , those FPSs of deltas can be the difference between slowdowns and jumps and a perfect fluid gaming experience.
Believe I have tested doom 3 in an AXP @2GHz and in an A64 3200+ (@2GHz). Even if If I used a faster videocard in the AXP the action was a little jumpy when lot of monsters in screen, with the A64 the Action was perfectly smooth even with a slower videocard, both systems had 1GB of RAM. That is what I mean. I am not saying it's not playable, it is perfectly playable in the AXP, but you can certainly see the difference with the A64 while gaming.
So in those Real world tests you will see the difference.
This have been discussed many times before.
Also you are only based on gaming tests, again, what with the rest of the tests where are Real apps and the AXP gets also whipped.
If you can't notice the difference between AXP and AMD64 CPUs in the apps you use, that's you, but don't come to us saying that the difference is not noticeable, most of us have used AXP CPUs and AMD64 CPUs for years and we know for sure what the difference in performance is , and it's HUGE. The fact that you are not able to see the difference doesn't mean it doesn't exist and isn't HUGE, because all of us who have tested and used both plataforms for a long time will confirm there's a singnificant difference, Also Anandtech Crew who for sure have more idea of computing testing than you.
You just don't give up. Just accept it. clock for clock AMD64 CPUs blow away Athlons XP (talking about pure CPU power) as I have showed in all the benchies. There is nothing more to add.
If you say again that the opteron 939 @3GHz is not noticeable faster than the AXP-M @2.2GHz I will quote you in my sign (That's a laughable statement:laugh::laugh
. You may understand that cannot be good for your reputation in the forums.
Let's not forget that if you are using a socket A platform and want to upgrade your agp video card to something more substantial, it is gonna cost. This is why I threw in the towel for socket A and switched to socket 754. PCI video cards especially in the mid range are dirt cheap.
A new socket 754 combo with a Sempron is actually cheaper then a socket A Barton combo now. It does have a noticable performance improvement (not earth shaking) but when it is time to buy a video card you can get a whole lot more PCI-E then AGP for the same money.
A new socket 754 combo with a Sempron is actually cheaper then a socket A Barton combo now. It does have a noticable performance improvement (not earth shaking) but when it is time to buy a video card you can get a whole lot more PCI-E then AGP for the same money.
Originally posted by: carlosd
If you say again that the opteron 939 @3GHz is not noticeable faster than the AXP-M @2.2GHz I will quote you in my sign (That's a laughable statement:laugh::laugh
Here ya go, all summed up as best as possible to fit in one little line:
In many instances, a budget Opteron system @3GHz will seem no faster than a budget AXP-M system @2.2GHz.
This is a thread about budget systems, and each has their own idea of what that might be. (just like each has their own idea of faster and blown away). Like rogue1979 stated, one of socket A's biggest problems is AGP. To invest in a high end AGP card is a waste. With PCIe, you can get a card like ATi x800 for $105 after rebate, which works great in a budget system. (although comparable AGP is now priced fairly close).
Eventually, even those hanging on to Socket A will need tp upgrade, and it is doubtful that they will be able to move an AGP card to the next gen motherboards/CPU's that they've been waiting for to get that WOW upgrade feeling.
For general Windows programs, the difference between the very old 1.4ghz T-bird in my retro game system and my Opteron at 3ghz is somewhat noticeable but still fairly small. I think if you optimize Windows effectively (getting rid of unnecessary GUI extras, background services and so on), old processors can keep up quite well with the newer ones there.
There are actually some situations where the XP is faster. I don't understand it, but with the same clock speed, cache size and memory, my old XP was definitely faster with Mathematica calculations. The difference was fairly small but consistently there in everything I tried. The XP could also go a bit higher on Dosbox CPU cycles without stuttering. I posted about this here a while ago but didn't get many responses. That's not to say I would stick with the XP just for that, as the 64 can be clocked much higher and will surpass it just due to that, but it's worth bringing up again. The XP may have had a slightly better FPU, lower cache latency or something like that.
There are actually some situations where the XP is faster. I don't understand it, but with the same clock speed, cache size and memory, my old XP was definitely faster with Mathematica calculations. The difference was fairly small but consistently there in everything I tried. The XP could also go a bit higher on Dosbox CPU cycles without stuttering. I posted about this here a while ago but didn't get many responses. That's not to say I would stick with the XP just for that, as the 64 can be clocked much higher and will surpass it just due to that, but it's worth bringing up again. The XP may have had a slightly better FPU, lower cache latency or something like that.
Zap
Elite Member
- Oct 13, 1999
- 22,377
- 7
- 81
QFT. After all, how much faster will web pages display with a faster CPU? How much faster will Solitaire bounce the cards with a faster CPU? How much faster will your music CD burn with a faster CPU?
My ~$100 video card choice would be those 6600GT PCIe that have been showing up for that price after rebate. Now if I'd only get my rebate.
I also have (just to be fair) an X800GTO 16 pipe modded, $145AR. Also waiting for rebate. :disgust: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: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter