The added cache to the 6xx series did little for performance, the cache latency got worse, and cedar mill is a die shrunk 6xx series, and presler is 2 of those slapped together. It has the same slower cache of the 6xx series, since no changes were made to the architecture at all, it's just a die shrink, so again, your numbers just don't add up. Presler will be better than the 8xx series for sure, but until they are widely available, you don't know how well they will overclock on average, and what the temps will be like. Like I said, my 830 runs almost as hot at stock speeds on water cooling, as my X2 with the stock cooler, not even using AS5, and while presler does run a little cooler, I doubt it will be that much cooler, it's a die shrunk prescott after all..Presler DOES have potential, but it's not released yet, so you can't give a valid comparison no matter how much you insist on your single synthetic benchmark being so great, and make points you can not prove.
The added cache to the 6xx series did little for performance, the cache latency got worse, and cedar mill is a die shrunk 6xx series, and presler is 2 of those slapped together. It has the same slower cache of the 6xx series, since no changes were made to the architecture at all, it's just a die shrink, so again, your numbers just don't add up. Presler will be better than the 8xx series for sure, but until they are widely available, you don't know how well they will overclock on average, and what the temps will be like. Like I said, my 830 runs almost as hot at stock speeds on water cooling, as my X2 with the stock cooler, not even using AS5, and while presler does run a little cooler, I doubt it will be that much cooler, it's a die shrunk prescott after all..Presler DOES have potential, but it's not released yet, so you can't give a valid comparison no matter how much you insist on your single synthetic benchmark being so great, and make points you can not prove.
CPU2000 rates test becomes my single synthetic benchmark ?
SPEC CPU test is very important performance test of CPU.
In the world, almost all kind of CPU (include Intel/AMD/IBM/SUN/HP ect.) join into this important test.
So we can see how powerful the integer/FP performance of the CPU.
If you can find another more important test, handred kinds of CPU from those most important processor company will join.
Please tell me.
The L2 latency of PD8xx may be higher than Prescott a little, but nobody says the L2 latency of Presler is higher than PD8xx. We need test to prove it.
65nm Presler is better than the 90nm 8xx series, and the fact from INQ has shown that they are much more easy to overclock.
A lot of people also believe it.
Please tell me.
The L2 latency of PD8xx may be higher than Prescott a little, but nobody says the L2 latency of Presler is higher than PD8xx. We need test to prove it.
65nm Presler is better than the 90nm 8xx series, and the fact from INQ has shown that they are much more easy to overclock.
A lot of people also believe it.
/slap Betwon
gg...
Synthetic benchmarks mean nothing, if you rely on them to prove your point you are not very bright. Check out the real world performance tests anand and many others have performed...
To get back on topic: the man states HE DOES NOT want to OC at all!! Go with either AMD solution and u will be more than happy, reuse your ram and it won't hinder your performance more than a a couple %. If you were to go with the pentium solution you would need some quality DDR2 to get the system running full speed, which means more $$ spent.
gg...
Synthetic benchmarks mean nothing, if you rely on them to prove your point you are not very bright. Check out the real world performance tests anand and many others have performed...
To get back on topic: the man states HE DOES NOT want to OC at all!! Go with either AMD solution and u will be more than happy, reuse your ram and it won't hinder your performance more than a a couple %. If you were to go with the pentium solution you would need some quality DDR2 to get the system running full speed, which means more $$ spent.
SPEC CPU test is different with "Synthetic benchmarks" such as PCMARK/SYSMARK etc.
That is a very important test, because it can really show the integer performance and float point performance of CPU.
So Power5/Itanium2/K7/K8/P3/P4/PM...., all join into this test. It can give you the real answer about the performance of CPU.
Why do you hate spec cpu2000? It is much more scientific than many other test to show the performance of CPU.
Guys, if you are looking for value don't take OC's into consideration as they are not guaranteed. I have had 3 Opteron 175's one OC'd very well, one ok, the and other one poorly. It's kinda like saying, well I'm going to take x job even though it pays less because I can get overtime and will make more money anyway. Maybe, if the overtime is there to get but you'll have to work more to get the OT pay just to make up that ground.
Yeah, if you get a cheap proc and it OC's like mad you got good value but the vaue was determined AFTER the purchase and not before. There are a lot of folks who bought 144's and got tremendous OC's and others got little and are bitter because they had an expectation. I guess what I'm getting at is don't lowball you purchase and expect great things and then be bummed about it later because it didn't work out the way you wanted.
Yeah, if you get a cheap proc and it OC's like mad you got good value but the vaue was determined AFTER the purchase and not before. There are a lot of folks who bought 144's and got tremendous OC's and others got little and are bitter because they had an expectation. I guess what I'm getting at is don't lowball you purchase and expect great things and then be bummed about it later because it didn't work out the way you wanted.
Some people bias the CPU's OC:
They think that:
If you meet AMD CPU that can not OC==You are lack of luck.
If you meet Intel CPU that can not OC==You are cheated.
They think that:
If you meet AMD CPU that can not OC==You are lack of luck.
If you meet Intel CPU that can not OC==You are cheated.
Markfw
Moderator Emeritus, Elite Member
- May 16, 2002
- 27,241
- 16,107
- 136
I didn't see a post from the OP thats says he won't OC, and I have 3 X2 3800's, and the worst of the bunch does 2500.
betwon, until you can buy a pressler and we see the price, and a benchmark review from anandtech or other reputable site, stop quoting synthetic benchmarks that can't even be verified and are crap to start with.
betwon, until you can buy a pressler and we see the price, and a benchmark review from anandtech or other reputable site, stop quoting synthetic benchmarks that can't even be verified and are crap to start with.
A boy.
The facts about SPEC CPU2000 let you feel angry?
What do you think about the most important performance test which can show the interge/FP performance?
Perhaps an even better question is, why would you want to know FP/integer performance instead of the performance of the chip in real applications?.....
Presler will probably perform about the same as the current P D's, sometimes a tad better, sometimes a tad worse, at same clock (judging from prescott 2M).
Presler will probably perform about the same as the current P D's, sometimes a tad better, sometimes a tad worse, at same clock (judging from prescott 2M).
Someone always try to detemine FP/integer performance by some test such as games or and so on.
There is the much more scientific test--spec cpu test.
In real applications of each people, PD820 witout OC maybe defeat opteron@2.4G, or opteron@2G maybe defeat PD@4.5G.
There is a joke:
"I find the gound is plain , Who knows that the earth is a ball? He must try to cheat me."
In real world applications, even the 840 extreme edition loses to the 3800+ X2 in most applications. I don't care how well it scores in a POS synthetic application no matter what it measures, because all it is is synthetic, and does not show how the CPU really performs. I own both an X2, and a Pentium-D, and the pentium-d loses in everything, while running as hot on water at stock speeds, as my overclocked X2 on the stock heatsink. You can spout off your scientific crap all you want, but for programs that actualy get used, the pentium-d doesn't come close. Video encoding for example, something people actualy use their CPU for, if you think people buy a proccessor to sit around doing nothing but running synthetic benchmarks, you really need to get a clue. And all you can do with Presler is speculate, because it isn't out yet. All benchmarks are for is to compare similar proccessors, to see if they are getting similar numbers, so you know everything is working properly, they have no bearing whatsoever on real world applications.
The performance of PD840EE is better than 3800+ X2 in most applications.
you own both an X2, and a Pentium-D, and the pentium-d loses in everything. It is your mistake.
No one who is straight will say that PD will lose in everything, of course you maybe:
"I find the gound is plain , Who says that the earth is a ball? He must try to cheat me."
Markfw
Moderator Emeritus, Elite Member
- May 16, 2002
- 27,241
- 16,107
- 136
So you are calling both of us a liar ? Where are your personal benchmarks from your own systems so you can say that ? Oh, don't have any ? just synthetic benchmarks from the Inquirer ? I guess its you that made a mistake by posting.
FelixDeCat
Lifer
- Aug 4, 2000
- 30,973
- 2,676
- 126
I agree with you Betwon, Intel has long made quality chips for decades. Im quite satisfied with my Intel solution and dont plan to upgrade any time soon. I refuse to AMD. Its rubbish IMO.
The thing about AMD users, they change chips like they change underwear. You can be satisfied with your Intel product for a year or more.
Zebo
Elite Member
- Jul 29, 2001
- 39,398
- 19
- 81
From Anandtech: The performance of PD 820@2.8G is about equal to Athlonx2@1.8G.
From SPEC CPU: The performance of PD 820@2.8G is about equal to Athlonx2@2.0G or more.
From SPEC CPU: The performance of PD 820@2.8G is about equal to Athlonx2@2.0G or more.
I have also been struggling with the decision whether to buy Pentium D or AthlonX2 based sytem. I was first going into Athlon X2 but then turned to Pentium D because I saw it more sustainable investion. I think I'll buy some cheap 775-processor for time being and go for 65nm's when they come out. BUT now when I read this thread I came doubtlful again.
My main use for the sytem would be music production (Sonar 5, 64bit and multithreaded) and video editing. (no gaming, no need for SLI etc.)
NO Overclocking!
I was thinking of Intel stuff because there will be dual cores with HT enabled (and 840 EE ofcourse but I will not buy that) and though of having 4 threads running just for Sonar is appealing.
What does the AMD's roadmap look like? what does the AMD have to offer a year from now?
One more very important question:
If I buy, for example Gigabyte GA-8I955X Royal or some Asus 955x, will they be compatible with the forthcoming 65nm processors? I understand that they are, but it has been quite difficult to find a definiteve yes or no for an answer.
Thänks!
-Tomi
My main use for the sytem would be music production (Sonar 5, 64bit and multithreaded) and video editing. (no gaming, no need for SLI etc.)
NO Overclocking!
I was thinking of Intel stuff because there will be dual cores with HT enabled (and 840 EE ofcourse but I will not buy that) and though of having 4 threads running just for Sonar is appealing.
What does the AMD's roadmap look like? what does the AMD have to offer a year from now?
One more very important question:
If I buy, for example Gigabyte GA-8I955X Royal or some Asus 955x, will they be compatible with the forthcoming 65nm processors? I understand that they are, but it has been quite difficult to find a definiteve yes or no for an answer.
Thänks!
-Tomi
Many successful tests of presler had been made. It used:
Intel Platform Gigabyte GA-8I955X Royal (Socket 775), Rev. 1.1 Intel 955X Chipset, BIOS F8
The X2 is my first AMD chip since the K6-2. If you really think AMD is rubbish, you are behind the times. My X2 has given me no problems at all. My prescott based pentium 4's run very hot, and my 3.4 was throttling at stock speeds with an XP-120. Not being able to run at it's rated speed without overheating seems like a problem to me..my Northwoods never game me trouble, but Prescotts are space heaters and slower than the older northwood in several things.
My pentium-d was just below throttling at stock speeds, slower than my single cores in non-smp aware apps, and destroyed by the X2 in everything. Now that it's water cooled, it runs at about the same temp as my overclocked X2 on the stock AMD cooler. The Smithfield CPU's are running at the low end of Intels clockspeeds because of the heat issues.
My Pentium-M has given me no trouble either, it runs cool, even overclocked with the dinky little cooler from the CT-479 adaptor. It's not a perfect CPU, but it beats the pentium-4's in a lot of things, and is just slightly behind the X2 in many(non-SMP aware) things at the same clock speed.
IF Yonah can overclock like Dothan does, and is reasonably priced, then I think it will be very competative. IF Presler can overclock on air as well as it's rumored to, it will be very competative. But Smithfield just plane sucks compared to the X2. If you refuse to use the cooler running, but faster dual core, thats your loss.
The only dual cores that will have hyperthreading are the extreme editions, so you won't find an inexpensive dual core with hyperthreading. The other problem reported with the extreme edition, is that windows see's 4 logical cores, it doesn't know which are the real cores, and which are the hyperthreading..so it actualy ends up performing worse..because a program will use a real core, and hyperthreading, rather than using the 2 real cores. If you aren't going to overclock, it doesn't even make sense to wait for the 65nm dual cores, since they will only be going to 3.4ghz, and have no architectural improvements. Cedar Mill= die shrunk 6xx series prescott, and Presler = 2 cedar mill die packaged together. Overclocked, the Presler COULD do well, at stock speeds, the X2 is already ahead.
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