![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Anandtech has reviews of the AMD 64 X2 on their website.
- Thread starter wilson502
- Start date
impressive, expecially because their 'emulated' a64 x2 should be a tad slower than the actual one - i hope some sites put up some maya benchmarks and not just max.
steept pricing on the opteron front, but then again that had to be expected: the ability to use existing infrastructure paired with fact of being roughly one year on the market before intel is, is most certainly still making this the cheaper upgrade path.
steept pricing on the opteron front, but then again that had to be expected: the ability to use existing infrastructure paired with fact of being roughly one year on the market before intel is, is most certainly still making this the cheaper upgrade path.
SlingXShot
Limp Gawd
- Joined
- Dec 25, 2004
- Messages
- 152
Is the 4800 processor releasing with the 4400+
I know the 4800 would be on top of all benchmarks, the 4400 is slightly in the middle-top if you average.
I know the 4800 would be on top of all benchmarks, the 4400 is slightly in the middle-top if you average.
SlingXShot said:
"That being said, we have made it a point to bring you a preview of Athlon 64 X2 performance in this article, despite the fact that AMD isnt introducing the chips for another two months."
sounds like they are going to release all at the same time IMHO. 4400+ seems to be the sweet spot when it comes to bang/buck considerations. the premium they charge on the 4800+ seems a bit high
SlingXShot
Limp Gawd
- Joined
- Dec 25, 2004
- Messages
- 152
wizzackr said:"That being said, we have made it a point to bring you a preview of Athlon 64 X2 performance in this article, despite the fact that AMD isnt introducing the chips for another two months."
sounds like they are going to release all at the same time IMHO. 4400+ seems to be the sweet spot when it comes to bang/buck considerations. the premium they charge on the 4800+ seems a bit high
2 months + 1 month = 3 months
mpcamer1220
[H]ard|Gawd
- Joined
- Oct 2, 2003
- Messages
- 1,717
i want one
Sure their sweet, but for the first time in years i'm thinking I might go intel on my next upgrade. If i can get a dual core p4d for [email protected] when the very cheapest (and probably better performing than intels highest end dualcore) is over 500$, that makes alot more sense to me. It seems to me AMD is freaking out, scared that their dual cores athlons will stunt opteron sales and their gonna stop it by pricing them into rediculousness. It never ceases to amaze me how AMD can snatch defeat from the jaws of victory like they do over and over again. I love AMD but sometimes they make me wonder about their logic.
7718 said:
"Remember that at 90nm, AMD can produce around half as many dual core CPUs as they can single core CPUs per wafer - so they need to be very careful about demand. You will notice later on in this article that AMDs strategy involves keeping prices higher and introducing lower quantity CPUs first, in order to ensure that their single core CPUs still have a market and that they arent committing to more that they can deliver. At the end of the day, AMD is still a much smaller manufacturer than Intel and thus they have to play their cards very differently" - anandtechs'
although i feel your pain - AMDs pricing scheme does make sense, so i wouldn't say AMD "snatched defeat from the jaws of victory". we'll have to wait and see how pricing is going to develop in the real world once the chips are out on the market. Now that intel's HT advantage seems to be gone i think i'll switch back to AMD this time - i'm just waiting to see which one overclocks better
eloj said:
How is that not serious? None of the computer people I know (even ones who run somewhat popular websites) use Linux even for their servers. In the computer world it really is almost all Windows(TM) , Windows(TM) and well Windows(TM).
What a trip, how people react differently to news. I consider myself more or less AMDcentric , I've been buying their products since my first socket A duron 650mhz chip back in tha day, but the whole dual core pricing thing might push me to intel. You have an intel sys and your flip floping the other way.. Did I wake up in Bizzaro world?
7718 said:
well, as is said before - i'm not sure, yet. the fact taht i have an intel system doesn't help me either, as i'm going to have to upgrade the entire rig anyways. it all depends on which ever processor is going to preform better in rendering maya files, which one overclocks better, if the gap in between the two regarding prices is going to be that huge in the end, if i can get hold of a friends s939 system for little money, i'm not going to buy the new system right when it comes out giving prices a little time to settle etc. etc.
...PLUS i wanted to toy around with an A64 for ages, so i'd definitely pay a price premium just for the fun of it, as long as it stays in reasonable limits.
tsuehpsyde
Supreme [H]ardness
- Joined
- Oct 22, 2004
- Messages
- 6,604
I've never been so excited to read a review before 4800+ Oh man I want that 4.8 GHz and 2 megs of cache worth of AMD power in one chip!
4800+ Oh man I want that 4.8 GHz and 2 megs of cache worth of AMD power in one chip! SlingXShot
Limp Gawd
- Joined
- Dec 25, 2004
- Messages
- 152
tsuehpsyde said:I've never been so excited to read a review before
4.8ghz Intel you mean.
4800+... I wonder how a thunderbird core athlon would perform at 4.8ghz.. I wonder if AMD manufactures one tbird everytime they refine their manufacturing process so they can use it to make sure their model numbering sys is still correct. Anybody have a high end nf2 board w/phase change cooling they want to OC the snot out of so we can see some serious comparisons? AXIA chips were capable of 1500ish mhz with proper cooling right? I don't know if anybody carried one of those chips over to a mature socket a board or not. Somewhere somebodys ocing a oldschool socket A athlon somewhere right?
tsuehpsyde
Supreme [H]ardness
- Joined
- Oct 22, 2004
- Messages
- 6,604
4800+ = 2x 2.4GHz core.....2x 2.4 = 4.8 so no, i meant AMD GHz.
(cf)Eclipse
Freelance Overclocker
- Joined
- Feb 18, 2003
- Messages
- 30,027
yup, and then they drop 200pr for 1/2 the cache
tsuehpsyde
Supreme [H]ardness
- Joined
- Oct 22, 2004
- Messages
- 6,604
The benches are crazy! Things Intel used to be the king of it's getting stomped out of by AMD....and they're using the middle chip for AMD (4400+) while it's going against the new EE with Hyperthreading (4 cores, in a manner of speaking) and an FX-55 (obviously not cheap or slo) and it's slaying left and right. I don't wanna sound like a fan-boy but hot damn!
Hito Bahadur
[H]ard|DCer of the Month - December 2006
- Joined
- Jul 8, 2004
- Messages
- 3,169
tsuehpsyde said:The benches are crazy! Things Intel used to be the king of it's getting stomped out of by AMD....and they're using the middle chip for AMD (4400+) while it's going against the new EE with Hyperthreading (4 cores, in a manner of speaking) and an FX-55 (obviously not cheap or slo) and it's slaying left and right. I don't wanna sound like a fan-boy but hot damn!
For a CPU that will be almost 1/2 the price it's performance isn't too bad at all.
With respect to pricing. I think AMD understands that they are going to be manufacturing limited and the margin in the server world is much greater. From a business standpoint, they need to ensure they grab as much market share as possible in the next 8-12 months before the dual core Xeons come out as that will ensure a longer term, higher margin, market share. Unfortunately that means significantly higher prices for the non-server market. I would think prices will get much more competitive when Fab 36 comes up and they have switched over to 300mm wafers.
/dev/null
[H]F Junkie
- Joined
- Mar 31, 2001
- Messages
- 15,182
eloj said:
Word to that. It would be nice to see kernel compile times or something similarly useful to those of us who aren't graphics artists and want to use this for work purposes.
Rob
Anands "Final words" said:
AMD Marketing department said:
Well, possibly a bit harsh, but starting his concluding words with a misleading statement like that makes me a little uneasy.
I use my PC pretty much just for gaming, surfing/research and E-mail, and I suspect that I'm far from alone in that.
Money.Down.Toilet.
7718 said:
The whole PR system no longer makes sense. The new dual cores don't compare to a 1.4 T-bird at 4.8 GHz (official) or a 4.8 GHz prescott. Neither even exist. The PR system is now perpetuating the very myth (GHz matters) that it was originally designed to combat. Just give the damn things a model number like the Intels, the Opterons, and the FXs, and be done with it.
Note also, that besides the 2T timing and ECC memory, the motherboard was designed with 2cpu in mind. Anandtech was lucky to even have one cpu running with the current bios. This will probably impose a performance hit also.
A real dual-core 4400+ will IMHO be 10% faster.
A real dual-core 4400+ will IMHO be 10% faster.
I find it strange that the 2.2GHz, 1MB cache per core DC has a PR number that is 200 lower than the 2.4GHz, 512KB cache per core DC (4400+ vs 4600+). By AMD tradition, these two should perform close to each other and should have a PR number within 100 of each other (like the Venice 3800+ and San Diegn 3700+).
Yes, but the cooling for that Pentium D will make up for the difference7718 said:
All kidding aside, the overall costs to operate a Pentium D significantly offset the lower purchase cost of the processor alone. With a Pentium D setup, one needs ensure that the chassis, power supply, motherboard, and CPU cooler are up to the task of operating these demanding CPUs.
Do you really believe that?
The Athlon 64 platform lacks many functions that server and workstation operators can typically find on the Opteron platform--ECC, registered DDR, PCI-X, SCSI controllers, hardware monitoring DSPs, etc. I wonder how many of these operators are tempted to sacrifice these things to cut costs.
I refuse to be AMD's PR spokesman, but I think you need to reconsider the implications of the statement that they are "freaking out."
Jonsey said:
AMD is going to use intel's own weapon against them for as long as they can.
They probably will switch in the future since it wouldn't make sense to have the athlon 64 154000+, but as long as it's working for them, they should keep using it.
xonik said:
They are causing more confusion by sticking to a system that no longer makes sense. Why would you want AMD to make sales from clueless customers? They should be doing their best to clarify processor performance, not make useless comparisons between a single core processor sold four years ago and a modern, dual core processor with a higher clock speed and IPC.
Jonsey said:
Are you actually serious about this question?
Best answer...
Because Intel has cornered the "clueless user" market for so long.
And sales are sales.
People don't want to read articles on IPC. They want a number to look at to compare all systems.
Intel banked on that fact and trained everyone to look at that number for a long time.
So when they shot themselves in the foot with the p4, and couldn't get any higher numbers, they went to the model clasification.
First off, the comparison to the "Thunderbird" Athlon has been dropped ever since the Athlon 64 was released. AMD themselves make no such distinction as to what the model number compares with. I believe that the model number was normalized to Pentium 4 processors as they progressed--with the exception of dual core processors, of course.
Anyways, that's beside the point. I don't care much for AMD's Athlon 64 system anyways, but that does not mean they should drop it. AMD can't afford to appear capricious right now. They should stick it out for a while before changing. Another thing: I don't care much for Intel's naming system either. The 800 series is not 33% faster than the 600 series, nor is it 14.3% faster than the 700 series (Pentium M). It's just a bad way to show relative performance differences.
The best way, in my opinion, would be something like this:
Pentium D 2.4D-2P or Pentium D 640-2P
or
Athlon 64 X2 3500+ or Athlon 64 X2 Model 175
In other words, give the processor some sort of performance indicator, followed by an indication of the number of cores present.
Anyways, that's beside the point. I don't care much for AMD's Athlon 64 system anyways, but that does not mean they should drop it. AMD can't afford to appear capricious right now. They should stick it out for a while before changing. Another thing: I don't care much for Intel's naming system either. The 800 series is not 33% faster than the 600 series, nor is it 14.3% faster than the 700 series (Pentium M). It's just a bad way to show relative performance differences.
The best way, in my opinion, would be something like this:
Pentium D 2.4D-2P or Pentium D 640-2P
or
Athlon 64 X2 3500+ or Athlon 64 X2 Model 175
In other words, give the processor some sort of performance indicator, followed by an indication of the number of cores present.
Steel Chicken
Supreme [H]ardness
- Joined
- Jan 29, 2001
- Messages
- 4,798
if these things are too expensive, maybe it will push down single core opteron prices....yummm
xonik said:
I never saw an announcement from AMD as to what their PR numbers stand for, but like you, I assume they stand against P4 performance. The question is, does a dual core AMD 4800+ have performance comparible to a prescott at 4.8 GHz? No. In single threaded benchmarks, I would imagine the prescott would win by a large margin. In multithreaded apps, I would imagine the AMD64 4800 would. I say imagine, because a 4.8 prescott doesn't exist, so who knows for sure. Besides, why use a PR system designed for a single core processor when the X2 will be competing with dual core parts from intel?
As for model numbers implying performance, that's only true to a point. A 800 series implies better performance than a 600 series, but not necessarily the ratio of 800/600. Is a BMW 5 series 1.6 times better than a 3 series? Is a FX-55 only 3.7 percent faster than a FX-53? Model numbers don't have to be a indicator of pure performance.
As far as when AMD should impliment new numbers, I would think that a major change in the way processors are marketed is as good a time as any. I read that there will be no more faster single core AMD64's. So why not start fresh and give the dual core AMD64's a new numbering scheme?
I like your numbering scheme, just take out the "Athlon 64" part and call it the X2-175.
Hito Bahadur
[H]ard|DCer of the Month - December 2006
- Joined
- Jul 8, 2004
- Messages
- 3,169
I'll never be able to find the reference to this, but my understanding was that the AMD PR rating was based on AMD products and never was based on Intel computers. It was basically the performance of Athlon XP's compared to the Speed of an Athlon.
"Their claim is that the PR rating on a chip denotes how fast an AMD Athlon Thunderbird processor would have to be to match the rated chip. "
http://linux.cudeso.be/amd-pr.php
This isn't the original place I remember seeing it a long time ago, but this is what I remember the reference being....
"Their claim is that the PR rating on a chip denotes how fast an AMD Athlon Thunderbird processor would have to be to match the rated chip. "
http://linux.cudeso.be/amd-pr.php
This isn't the original place I remember seeing it a long time ago, but this is what I remember the reference being....
Except games, but that's another story. Unfortunately for the marketing people, relative performance does not stay constant from application to application, which is one reason for all of this confusion.Jonsey said:
It doesn't seem right, at this moment. If multithreaded applications were ubiquitous at this point, I think you'd be singing another tune. In this reality, however, you're absolutely right.
True. Perhaps I oversimplified the nature of this problem.
Let's look at a real problem between processor lines. Let's look at the 600 series vs. the 700 series. The Pentium M 700 series often outperforms the Pentium 4 600 series counterpart, yet in other areas, it's significantly at a disadvantage. We see similar results when comparing the Pentium M 700 series with the Pentium D 800 series.
What, we can't compare mobile to desktop? Tell that to Joe Sixpack and see if he cares. See my point? Intel's numbering system is misleading, too.
Good point, I agree.
That's an old statement, and I don't even know if it's an official AMD statement anyways.Hito Bahadur said:
Hito Bahadur
[H]ard|DCer of the Month - December 2006
- Joined
- Jul 8, 2004
- Messages
- 3,169
xonik said:
Old doesn't mean wrong.
I don't dissagree with what you have said, but there needs to be some numbering system and I doubt there is one everyone could agree with.