Sign in with
Sign up | Sign in
Your question

65nm Single Core A64s Announced

Tags:
Last response: in CPUs
Share
Related resources
February 20, 2007 2:39:15 PM



Ok, then the next rational question is WHY?

Exactly what I was thinking. At those prices why not spend the little bit extra and get a dual core 3600+? Seems kinda pointless to me. They should sell these for less and rebrand them as Semprons.

Oh no! You can save all of nine whole dollars and get a singlecore 3800. That's gotta be faster! 200 whole whatevers! It's such a deal! 8O
February 20, 2007 2:42:00 PM

Quote:
Exactly what I was thinking. At those prices why not spend the little bit extra and get a dual core 3600+? Seems kinda pointless to me. They should sell these for less and rebrand them as Semprons.


You have made a good point :D 
February 20, 2007 2:47:39 PM

Quote:
You have made a good point :D 


Don't you ever wonder why we shouldn't all band together, get an investment bank or two behind us, and take over AMD? Then we could run it via THG committee. It's quite obvious that we know how to run that company far better than the current bunch of wankers. LET'S DO IT!!!!! :D 
February 20, 2007 2:54:26 PM

Quote:
You have made a good point :D 


Don't you ever wonder why we shouldn't all band together, get an investment bank or two behind us, and take over AMD? Then we could run it via THG committee. It's quite obvious that we know how to run that company far better than the current bunch of wankers. LET'S DO IT!!!!! :D 

US$8B 8O 8O
February 20, 2007 2:54:54 PM



Ok, then the next rational question is WHY?

Exactly what I was thinking. At those prices why not spend the little bit extra and get a dual core 3600+? Seems kinda pointless to me. They should sell these for less and rebrand them as Semprons.


Perhaps because if you're buying and running 1000 machines you save 20 x 1000 intialy and then save on energy as 45W is less than 65W.
February 20, 2007 2:58:23 PM

Quote:
Perhaps because if you're buying and running 1000 machines you save 20 x 1000 intialy and then save on energy as 45W is less than 65W.


I don't think the electric bill save of single-core A64s from dual-core A64s will be higher than 10W as most of the time the CPUs are idle.
February 20, 2007 3:15:15 PM

Quote:
Perhaps because if you're buying and running 1000 machines you save 20 x 1000 intialy and then save on energy as 45W is less than 65W.


I don't think the electric bill save of single-core A64s from dual-core A64s will be higher than 10W as most of the time the CPUs are idle.


So you think that a 65W chip idles better? If one is 45W how can it only be 10W lower than a 65W chip? With all things being equal it will always be 20W less. 50 of those is a KW-h.
February 20, 2007 3:20:01 PM

Quote:
So you think that a 65W chip idles better? If one is 45W how can it only be 10W lower than a 65W chip? With all things being equal it will always be 20W less. 50 of those is a KW-h.


I think you need to check the definition of AMD's TDP. :wink:
February 20, 2007 3:30:39 PM

Never ceases to amaze me that if one company does something its praised or ignored but if the rival does it someone has to rip them a new one.

Unless I'm mistaken, which I'm not, Intel are releasing SINGLE core versions of their new chips soon and no one has complained about this.

Currently there is still a market for single core processors. I would imagine that they are cheaper to build and can get higher yields. OEMS are happier to use these in their entry level systems as opposed to low end duals.

After all AMD/INTEL would not produce them if there was no market for them.

As BM said from a business perspective (buying 100's or 1000's) it makes sense.
February 20, 2007 3:32:32 PM

They can pump out a lot more single core chips per wafer than they can dual-cores. If people are buying them at only $10 less than a dual-core, they're making a lot more money from them than a dual-core processor. So they'll sell as many as they can.
February 20, 2007 3:59:32 PM

Quote:
So you think that a 65W chip idles better? If one is 45W how can it only be 10W lower than a 65W chip? With all things being equal it will always be 20W less. 50 of those is a KW-h.


I think you need to check the definition of AMD's TDP. :wink:


But the TDP will be rated the same. If one is rated at 45W it's 20W less than one rated at 65W, unless AMD is usng that new math.
February 20, 2007 4:00:36 PM

Quote:
They can pump out a lot more single core chips per wafer than they can dual-cores. If people are buying them at only $10 less than a dual-core, they're making a lot more money from them than a dual-core processor. So they'll sell as many as they can.


But from the performance-per-dollar view, Celeron D (Cedar Mill) family crushes Sempron family.
February 20, 2007 4:04:49 PM

Quote:
So you think that a 65W chip idles better? If one is 45W how can it only be 10W lower than a 65W chip? With all things being equal it will always be 20W less. 50 of those is a KW-h.


I think you need to check the definition of AMD's TDP. :wink:


But the TDP will be rated the same. If one is rated at 45W it's 20W less than one rated at 65W, unless AMD is usng that new math.

The TDP definition by AMD is the "worst maximum power dissipation" which is not easily achieved. Also most of the time the machines are in idle position. The power saving of getting 65nm processor instead of 90nm is not huge.

Look at the graph by X-Bit Labs:
http://www.xbitlabs.com/articles/cpu/display/athlon64-b...



Even for dual-core processor, the difference of power consumption numbers are marginal at idle. So the difference in power consumption would not be high.
February 20, 2007 4:17:17 PM

Quote:
They can pump out a lot more single core chips per wafer than they can dual-cores. If people are buying them at only $10 less than a dual-core, they're making a lot more money from them than a dual-core processor. So they'll sell as many as they can.


But from the performance-per-dollar view, Celeron D (Cedar Mill) family crushes Sempron family.

I'm talking A64s, 3200+, 3500+, 3800+. The 3800+ @ $88 ($92 on Newegg) massacres P4s and even many PDs with ease, while costing only ~$30 more than Celeron Ds.

The $65 A64 3000+ also pile drives almost any Celeron D you want to try and match it against.

Trying to say Semprons are what Celeron Ds would really have to match up against is silly.
February 20, 2007 4:19:13 PM

Quote:
I'm talking A64s, 3200+, 3500+, 3800+. The 3800+ @ $88 ($92 on Newegg) massacres P4s and even many PDs with ease, while costing only ~$30 more than Celeron Ds.

The $65 A64 3000+ also pile drives almost any Celeron D you want to try and match it against.

Trying to say Semprons are what Celeron Ds would really have to match up against is silly.


I cannot give the URL at this time, but I have saw a benchmark that A64 3K+ is crushed by C-D 352.
February 20, 2007 4:21:16 PM

Quote:
I'm talking A64s, 3200+, 3500+, 3800+. The 3800+ @ $88 ($92 on Newegg) massacres P4s and even many PDs with ease, while costing only ~$30 more than Celeron Ds.

The $65 A64 3000+ also pile drives almost any Celeron D you want to try and match it against.

Trying to say Semprons are what Celeron Ds would really have to match up against is silly.


I cannot give the URL at this time, but I have saw a benchmark that A64 3K+ is crushed by C-D 352.

Find it if you can, I'd like to see.

I just noticed that the Sempron 3600+ is equally priced to the A64 4000+ (the Sempron is out of stock). That's just silly. It's like a free 400Mhz, double the cache and higher HT speed.
February 20, 2007 5:24:35 PM

I think the AMD pricing department must be staffed by Rastas who think today is Marley's birthday.
February 20, 2007 7:34:14 PM

Quote:
So you think that a 65W chip idles better? If one is 45W how can it only be 10W lower than a 65W chip? With all things being equal it will always be 20W less. 50 of those is a KW-h.


I think you need to check the definition of AMD's TDP. :wink:


But the TDP will be rated the same. If one is rated at 45W it's 20W less than one rated at 65W, unless AMD is usng that new math.

The TDP definition by AMD is the "worst maximum power dissipation" which is not easily achieved. Also most of the time the machines are in idle position. The power saving of getting 65nm processor instead of 90nm is not huge.

Look at the graph by X-Bit Labs:
http://www.xbitlabs.com/articles/cpu/display/athlon64-b...



Even for dual-core processor, the difference of power consumption numbers are marginal at idle. So the difference in power consumption would not be high.


Wow, where are you guys coming from?

65W - 45W = 20W. Period.

No matter what you do or say that will always be the case. You are comparing chips that are actually rated the same TDP with lower clocks. I'm talking about a lower TDP regardless of clock speed.

I guess if you grab THGs guide and check single vs dual core power reqs you can then compare.
February 20, 2007 7:41:28 PM

Quote:
Wow, where are you guys coming from?

65W - 45W = 20W. Period.

No matter what you do or say that will always be the case. You are comparing chips that are actually rated the same TDP with lower clocks. I'm talking about a lower TDP regardless of clock speed.

I guess if you grab THGs guide and check single vs dual core power reqs you can then compare.


I'm emphatically not flaming you, but please help me to understand. Are you saying:

A 65W CPU runs at 65W at 100% usage.
A 45W CPU runs at 45W at 100% usage.

A 65W CPU runs at approx. 10W at 1% usage.
A 45W CPU runs at approx. -10W at 1% usage.

Does that mean that if I run 10 45W CPUs I am actually generating enough energy to run a 100W bulb?

:?
February 20, 2007 7:56:06 PM

Quote:
I think the AMD pricing department must be staffed by Rastas who think today is Marley's birthday.


Though I hate it I think they're going for Intel's throat. If Kuma is 20% faster than Core 2 then AMD can change 20%(not exactly) more or they could just intro at slightly lower and drop the bottom out from under Core2.

They have said in an interview that Barcelona will be priced the way the Dual Opteron was at intro.

I figure the same will happen with X2/X4/FX.
February 20, 2007 9:26:55 PM



Ok, then the next rational question is WHY?

Just a thought! :wink:
Could it be that for all but (7%)of the computer desktop market could run sufficiently well with a single core processor. No need for core 2's or X2's capable of .5 teraflops necessary?
Most of the gamers (2%) of the CPU market will want the high end equipment.
For the rest Single core processors are sufficient.

__________________
In answer to that eternal question, Booze is the answer. But now I don't remember the question. Everyone needs belief in something. I believe I'll have another beer!
February 20, 2007 9:31:36 PM

Quote:
You have made a good point :D 


Don't you ever wonder why we shouldn't all band together, get an investment bank or two behind us, and take over AMD? Then we could run it via THG committee. It's quite obvious that we know how to run that company far better than the current bunch of wankers. LET'S DO IT!!!!! :D 

OMG: THGAMD or AMDTHG :lol:  :lol: 

Now where is that Sucker Bankers Number!

__________________
In answer to that eternal question, Booze is the answer. But now I don't remember the question. Everyone needs belief in something. I believe I'll have another beer!
February 20, 2007 9:36:29 PM

I really like AMDTHG. It has a nice ring to it. If you get the investment banker on board, I'll design the logo! 8)
February 20, 2007 9:42:49 PM

I'm on it. :wink:

_________________
In answer to that eternal question, Booze is the answer. But now I don't remember the question. Everyone needs belief in something. I believe I'll have another beer!
February 20, 2007 10:02:08 PM

OK, report back to me when you have the $8 Bil. And come over and pick me up in our corporate jet. 8)
February 20, 2007 10:08:45 PM

You mean FUDMan can't fly?

Not that you -are- FUDMan, but I love your avatar.
February 20, 2007 10:09:55 PM

Quote:
You mean FUDMan can't fly?

Not that you -are- FUDMan, but I love your avatar.


Thanks for the props, dude. FUDMan can only fly when he is on the prowl for babes! 8)
February 20, 2007 11:39:34 PM

Quote:

Just a thought! :wink:
Could it be that for all but (7%)of the computer desktop market could run sufficiently well with a single core processor. No need for core 2's or X2's capable of .5 teraflops necessary?
Most of the gamers (2%) of the CPU market will want the high end equipment.
For the rest Single core processors are sufficient.


I agree with that. Also, there are many arguments on here about why people would chose an AMD over a C2D because of their superior performance over AMD. Well the simple answer is $$ matters to them more than anything. Most people aren't going to care if C2D beats out a cheaper AMD alternative by 500 points or 1:00 in some stupid benchmark. For those people, all they care about is it gets the job done and it saves them money.
February 21, 2007 12:21:37 AM

Quote:
Wow, where are you guys coming from?

65W - 45W = 20W. Period.

No matter what you do or say that will always be the case. You are comparing chips that are actually rated the same TDP with lower clocks. I'm talking about a lower TDP regardless of clock speed.

I guess if you grab THGs guide and check single vs dual core power reqs you can then compare.


I'm emphatically not flaming you, but please help me to understand. Are you saying:

A 65W CPU runs at 65W at 100% usage.
A 45W CPU runs at 45W at 100% usage.

A 65W CPU runs at approx. 10W at 1% usage.
A 45W CPU runs at approx. -10W at 1% usage.

Does that mean that if I run 10 45W CPUs I am actually generating enough energy to run a 100W bulb?

:?

Although this is not directly on point with the CPUs being discussed.....X-Bit compared AMD processers based on Winchester and Newcastle cores.
The power usage at full load had a TDP delta ranging from 22W at 1.8 GHz to 31 W at 2.2 GHz. At idle, the power delta was an almost constant 6 Watt difference.

Rated TDP definitely does NOT indicate what the power difference will be at idle. A 20W difference at load does not equal a 20W difference at idle.
I cannot think of a reason this trend would not hold true for the current generation of cores.

http://www.xbitlabs.com/articles/cpu/display/athlon64-9...
February 21, 2007 12:29:34 AM

Could it be that these single cores were manufactured to be dual core but one core doesn't work? So they sell them as singles rather than throw them away?
February 21, 2007 1:58:11 AM

Quote:
Could it be that these single cores were manufactured to be dual core but one core doesn't work? So they sell them as singles rather than throw them away?



I doubt it. Lima is a separate design following on from Manila single core. Just like Brisbane it is only a shrink to 65nm.
February 21, 2007 6:41:07 AM

Quote:

Although this is not directly on point with the CPUs being discussed.....X-Bit compared AMD processers based on Winchester and Newcastle cores.
The power usage at full load had a TDP delta ranging from 22W at 1.8 GHz to 31 W at 2.2 GHz. At idle, the power delta was an almost constant 6 Watt difference.

Rated TDP definitely does NOT indicate what the power difference will be at idle. A 20W difference at load does not equal a 20W difference at idle.
I cannot think of a reason this trend would not hold true for the current generation of cores.

http://www.xbitlabs.com/articles/cpu/display/athlon64-9...


OK, then (putting on his patented s***-disturber cap) then I take it that you are disagreeing with Baron's statement that the 20W or whateverW difference is constant all the way down to idle? (he says, running to find cover from the coming flamestorm)... :lol: 
February 21, 2007 10:47:05 AM

Quote:

Although this is not directly on point with the CPUs being discussed.....X-Bit compared AMD processers based on Winchester and Newcastle cores.
The power usage at full load had a TDP delta ranging from 22W at 1.8 GHz to 31 W at 2.2 GHz. At idle, the power delta was an almost constant 6 Watt difference.

Rated TDP definitely does NOT indicate what the power difference will be at idle. A 20W difference at load does not equal a 20W difference at idle.
I cannot think of a reason this trend would not hold true for the current generation of cores.

http://www.xbitlabs.com/articles/cpu/display/athlon64-9...


OK, then (putting on his patented s***-disturber cap) then I take it that you are disagreeing with Baron's statement that the 20W or whateverW difference is constant all the way down to idle? (he says, running to find cover from the coming flamestorm)... :lol: 

Yeah...although sometimes it is too easy (obviously!).
Here's the comparison I was talking about...

Newcastle 2.4 GHz (Full Load) - 81.6W
Winchester 2.4 GHz (Full Load) - 50.9W
Delta - 30.7W at full load

Newcastle 2.4GHz (Idle) - 19.4W
Winchester 2.4GHz (Idle) - 13.7W
Delta - 5.7W at idle

Not exactly a constant power differential!
February 21, 2007 11:52:48 AM

Quote:

Although this is not directly on point with the CPUs being discussed.....X-Bit compared AMD processers based on Winchester and Newcastle cores.
The power usage at full load had a TDP delta ranging from 22W at 1.8 GHz to 31 W at 2.2 GHz. At idle, the power delta was an almost constant 6 Watt difference.

Rated TDP definitely does NOT indicate what the power difference will be at idle. A 20W difference at load does not equal a 20W difference at idle.
I cannot think of a reason this trend would not hold true for the current generation of cores.

http://www.xbitlabs.com/articles/cpu/display/athlon64-9...


OK, then (putting on his patented s***-disturber cap) then I take it that you are disagreeing with Baron's statement that the 20W or whateverW difference is constant all the way down to idle? (he says, running to find cover from the coming flamestorm)... :lol: 

Yeah...although sometimes it is too easy (obviously!).
Here's the comparison I was talking about...

Newcastle 2.4 GHz (Full Load) - 81.6W
Winchester 2.4 GHz (Full Load) - 50.9W
Delta - 30.7W at full load

Newcastle 2.4GHz (Idle) - 19.4W
Winchester 2.4GHz (Idle) - 13.7W
Delta - 5.7W at idle

Not exactly a constant power differential!

Ok, so it is YOU (notice how I'm shifting all the responsibility and the targeting to YOU) who states that there is no way that a 45W CPU over its entire lifespan will use 20W less, consistently and equally, than a 65W CPU. Am I correct? :D 
February 21, 2007 12:14:12 PM

In Capt'n terms :1 vibrator jiggles at 1 gig draws for watts and two wiggles while same vibe jiggling at 10 gigs 20 wiggles and alot of jiggles got it?
February 21, 2007 12:26:54 PM

Quote:
In Capt'n terms :1 vibrator jiggles at 1 gig draws for watts and two wiggles while same vibe jiggling at 10 gigs 20 wiggles and alot of jiggles got it?


Jaydeejohn! It's so refreshing to finally have something explained in clear and simple layman's terms! :lol: 
February 21, 2007 12:34:36 PM

Its the old jigawatt theory, more bang for your buck sorta thing, but back to subject: being that both companies and trendy as well (ok truly needed) lowering wattage on cpu's as important as it currently is, has ANYONE done a full spec on usage? Say running certain apps for certain amounts of time and actually testing or comparing them in full use and at idle ?
February 21, 2007 1:55:33 PM

Quote:
Quote:
Quote:


...So you think that a 65W chip idles better? If one is 45W how can it only be 10W lower than a 65W chip? With all things being equal it will always be 20W less. 50 of those is a KW-h.


Baron, I think you might be mistaken here, as TDP for AMD in MAX wattage. It does not necessarily imply that either idle or even medium load states are lower by the same number of Watts. Because of how often CPU's are in an idle or medium load state, the power savings is likely to average less than 20W.
http://www.amd.com/us-en/assets/content_type/white_pape...
See page 12 or so of this document. The way I read it, a chip with the MAX TDP of 89W has an idle of 35W. Surely a 45W MAX TDP does not then have an idle of -9W.
February 21, 2007 2:36:13 PM

Quote:
Surely a 45W MAX TDP does not then have an idle of -9W.


My point exactly. It's like those useless gadgets that they try to sell bozos. Install this turboscrooeyquacker and save 20% of your gas! So if you install 6 of them, does your tank fill up and overflow? :lol: 
February 21, 2007 2:44:15 PM

Quote:
Surely a 45W MAX TDP does not then have an idle of -9W.


My point exactly. It's like those useless gadgets that they try to sell bozos. Install this turboscrooeyquacker and save 20% of your gas! So if you install 6 of them, does your tank fill up and overflow? :lol: 

My wife thinks that's how sales work if it's 1/3 % off a BOGO sale that's 83% off then. Nope it's 67%.
!