35W Athlon 3800+ X2

[needmorecarnitine]

http://www.fudzilla.com/index.php?option=com_content&task=view&id=8275&Itemid=1

supposedly available now at 33 euros. fudzilla says its 90nm?

I was thinking of waiting for the E5200 but if that doesn't come out soon and Gigabyte's GA-MA78GPM drops to around $80 before the Intel chip comes out, I might just pick up that combo

 

[Bao01]

Wow! This is the phantom chip AMD had promised 2 years ago. Never tried to buy one as it has always been expensive. Here's the one site that's supposedly still carrying it. $357

http://www.memory4less.com/m4l_itemdetail.asp?rid=fd_10&itemid=1439908571

If it lands at Newegg for 50 bucks, it might work well for an HTPC. But then, the E5200 is like 50% faster.

 

[pxc]

I looked it up too and found the same vendor. The 35W X2 EE was available for a short time, but it was very scarce and expensive back then.

It doesn't really make much sense for AMD to have Chartered Semi make those and sell it for peanuts, 90nm or not. The part number comes up on amdcompare.com as the 2 year old stepping and no newer 35W models are listed. AMD made no announcement and I'm thinking that European vendor either put the wrong part number/description up, or somehow acquired a surplus supply of the chips.

It would be nice for those to be produced again, but it doesn't look like it's a real product re-launch. Good deal for those who can order the chips. Grab 'em quickly, it's probably your only or last chance.

 

[needmorecarnitine]

it looks like it is a price comparison site and it lists six vendors that have it.

 

[Toaster Oven]

Call me skeptical but I bet if you downclock a 4050e to 2.0Ghz you'd see equal or better real world power consumption figures than that 90nm X2 3800+ EE.

 

[pxc]

it looks like it is a price comparison site and it lists six vendors that have it.

There's a few more listings than when the link was first posted.

I wonder if AMD was clearing out some closets in Dresden.
---

I thought the 4050e was only 2.1GHz. It's already a cut down 65W processor running at 1.15v-1.25v vs 1.30-1.375v on the other, faster rev G2 cores. I'd think the 4850e has a better chance of getting to 35W or below.

Lowering the 4050e (say @ 1.2v stock) to 2GHz @ 1.1v would drop the 45W TDP to ~38W. It may operate under 45W at stock speed and voltage, and the loads you put on it may never reach close to TDP anyways. Close enough. An underclocked 4850e in the same conditions as above would drop the 45W TDP to ~30W, if it can run stable at that speed and voltage. The price difference between the two models isn't much at newegg: $70 4050e vs $87 4850e.

The odd thing really is that AMD and Intel are going very separate routes with power figures. AMD is creating several tiers to look like it has lower power chips, and Intel is still sticking to what look like platform numbers. The results can be seen at places like xbitlabs in the isolated CPU power consumption tests. Lower end Intel 65W processors are using less power than AMD 45W processors. And while you can argue that Intel CPUs don't include a NB built in, the overall system power consumption also favors Intel. I think the only exception is the old X2 3600+ EE/BE-2300, but those weren't compared to newer 45nm processors.

 

[Toaster Oven]

I thought the 4050e was only 2.1GHz. It's already a cut down 65W processor running at 1.15v-1.25v vs 1.30-1.375v on the other, faster rev G2 cores. I'd think the 4850e has a better chance of getting to 35W or below.

Lowering the 4050e (say @ 1.2v stock) to 2GHz @ 1.1v would drop the 45W TDP to ~38W. It may operate under 45W at stock speed and voltage, and the loads you put on it may never reach close to TDP anyways. Close enough. An underclocked 4850e in the same conditions as above would drop the 45W TDP to ~30W, if it can run stable at that speed and voltage. The price difference between the two models isn't much at newegg: $70 4050e vs $87 4850e.

Sorry but your reasoning is faulty. TDP Rating != Real Power Consumption. Just because 2 processors fit under the same TDP rating envelope in no way means they have the same real world power consumption. A 4050e consumes less power under full load than a 4850e, been shown in many tests. And the difference in power consumption is almost a direct correlation between difference in operating speed, so the idea that a 4850e when downclocked will consume more than 20% less power than a 4050e at the same clock frequency is just wishful thinking on your part. Spending more on a 4850e and downclocking it to same speed as 4050e isn't gonna accomplish much besides lightening your wallet a bit more.

Anyhow, Tom's Hardware did a very nice test on power consumption of various AMD processors. Which confirmed my suspicion that 4050e (and even 4450e) sucks down less power than X2 3800+ EE which they also tested.

 

[pxc]

Sorry but your reasoning is faulty. TDP Rating != Real Power Consumption.

<insert jump to conclusions picture here>

It may operate under 45W at stock speed and voltage, and the loads you put on it may never reach close to TDP anyways.

Ahem, I already covered that and mentioned Intel 65W TDP CPUs often use less power than even AMD 45W CPUs. I used the word TDP deliberately to show how it would scale on that metric. Of course there are other factors like what I quoted from myself above. I was getting at TDP ratings are becoming more meaningless in the last paragraph I posted in #6.

The power consumption test in the THG post is just one sample of each anyways. The voltage range for each 4x50e model is 1.15v-1.25v, so power consumption can vary from what's on the graph. It's a small difference anyways, ~5W from highest 4x50e to lowest 4x50e and almost in line with the power predicted by frequency scaling alone (linear) and most of the difference could be from parts of the chip that don't scale with core speed (I/O, NB, etc).

 

[Toaster Oven]

<insert jump to conclusions picture here>

Ahem, I already covered that and mentioned Intel 65W TDP CPUs often use less power than even AMD 45W CPUs. I used the word TDP deliberately to show how it would scale on that metric. Of course there are other factors like what I quoted from myself above. I was getting at TDP ratings are becoming more meaningless in the last paragraph I posted in #6.

Yes you covered it on the Intel side and conveniently chose to forget it when debating the power consumption on AMD side? If it's becoming more meaningless like you said, why did you even try to make an argument on AMD power consumption based on TDP rating alone instead of real world figures?

The power consumption test in the THG post is just one sample of each anyways. The voltage range for each 4x50e model is 1.15v-1.25v, so power consumption can vary from what's on the graph. It's a small difference anyways, ~5W from highest 4x50e to lowest 4x50e and almost in line with the power predicted by frequency scaling alone (linear) and most of the difference could be from parts of the chip that don't scale with core speed (I/O, NB, etc).

So by that logic tests from any tech review site are unreliable because they only test with 1 sample unit per processor?

Given the same input voltage, sample variance between different chips is negligible. Both chips were tested at the same input voltage and they both share the same operational voltage range and architecture, so the difference in power consumption can be attributed to clock speed here. 5W isn't a small difference in a 45W envelope, that's more than a 10% difference in power consumption. And you are really grasping for straws with that last statement.

 

[pxc]

Yes you covered it on the Intel side and conveniently chose to forget it when debating the power consumption on AMD side?

No, I guess you didn't read it the first or second time. That section I quoted from myself was from the paragraph discussing AMD TDP scaling. It's still at post #6 if you ever find the time to read it. I gave that as an estimate with the caveat that it may use less power. If AMD has leakier devices that don't fit under 45W at 2.5GHz (4850e), they can sell those as 4450e or 4050e models. Sorry, but that's how it sometimes works.

So by that logic tests from any tech review site are unreliable because they only test with 1 sample unit per processor?

I didn't say that. I posted the reason why directly above, but in case you don't feel like reading what I posted i'll repeat: The chips have a voltage range and while you may purchase a chip with nearly identical power usage, you also may not. One sample of each model doesn't even give you an average.

5W over a 20% clock speed difference, considering how power increases linearly with frequency (all other things being equal), is a small range. Sorry that doesn't satisfy the strange conclusions you draw from skimming my posts, but what's there is there no matter how much you try to misinterpret it.

/plonk

 

[Toaster Oven]

No, I guess you didn't read it the first or second time. That section I quoted from myself was from the paragraph discussing AMD TDP scaling. It's still at post #6 if you ever find the time to read it. I gave that as an estimate with the caveat that it may use less power. If AMD has leakier devices that don't fit under 45W at 2.5GHz (4850e), they can sell those as 4450e or 4050e models. Sorry, but that's how it sometimes works.

No worries I read what you wrote. And what I'm saying is your estimate is worthless because in your example you assumed all the CPUs were operating at their max TDP rating. Which even you admitted is not true. Sure there may be a few samples here and there that are leakier than other, that's nothing new. But for your example to hold any relevance a 4050e and 4850e would have to share the same power consumption on average. Across all the reviews I've seen comparing both, that's not the case at all.

I didn't say that. I posted the reason why directly above, but in case you don't feel like reading what I posted i'll repeat: The chips have a voltage range and while you may purchase a chip with nearly identical power usage, you also may not. One sample of each model doesn't even give you an average.

No worries I read this too so lemme repeat what I wrote. Both chips share the same voltage range. Both chips will exhibit the same small differences in power consumption. But as 4050e is clocked lower, it will exhibit lower power consumption on average. If you are trying to convince me that 4050e does not consume less power than 4850e on average, then save your breath. There are plenty of reviews out on the web that show it to be the case.

5W over a 20% clock speed difference, considering how power increases linearly with frequency (all other things being equal), is a small range. Sorry that doesn't satisfy the strange conclusions you draw from skimming my posts, but what's there is there no matter how much you try to misinterpret it.

I love how you can convince yourself one thing while the numbers show another story altogether. Fail argument is still fail argument. Every test I've seen confirms 4050e power consumption is less than 4850e power consumption. No need for interpretation. How you jump from that to the conclusion that both processors have the same power consumption is beyond me

 

[ssjsongoku03]

what socekt is this proc?

 

[CptFalcon]

<insert jump to conclusions picture here>

Ahem, I already covered that and mentioned Intel 65W TDP CPUs often use less power than even AMD 45W CPUs. I used the word TDP deliberately to show how it would scale on that metric. Of course there are other factors like what I quoted from myself above. I was getting at TDP ratings are becoming more meaningless in the last paragraph I posted in #6.

The power consumption test in the THG post is just one sample of each anyways. The voltage range for each 4x50e model is 1.15v-1.25v, so power consumption can vary from what's on the graph. It's a small difference anyways, ~5W from highest 4x50e to lowest 4x50e and almost in line with the power predicted by frequency scaling alone (linear) and most of the difference could be from parts of the chip that don't scale with core speed (I/O, NB, etc).

Wait, how does Intel's C2D 65watt CPU use less energy than the AMD 4850e. Elaborate please.