why its reasonable for k8l to be q1

[cky2k6]

I just realized as possible strategy for why amd is probably releasing k8l q1 of 07, think about it, windows vista. The new os is probably going to bring quite a few people to upgrade their old dx9 systems, and plus vista is going to be 64bit and less legacy code oriented, which drops in perfectly with amd's plans on dropping some instructions from the x86 spec on their cpus, not recollecting if im confusing that with k10. in any case, its now alot more logical as to why amd isnt rushing out to push k8l out the door and encountering possible issues with either 65nm or zram or whatever. this kind of rings a bell with amd and intel's approaches... why i posted this in news, i dont know, move to amd section please.

[[XC] leviathan18]

vista is 32bits and 64bits

[Tim]

vista is 32bits and 64bits

Yea but only the stripped to the bone basic version is 32bits right? :)

[Dimitriman]

K8L needs to match Conroe for the whole 2007. Not just against the batch at launch in 2006.

[nn_step]

AMD is not going after desktops.. they are going after massive profit in the Server field.. Should good desktops come out of it.. so be it
See s940->s939

[Starscream]

Yea but only the stripped to the bone basic version is 32bits right? :)

The stripped down starter kit is 32bit only but all other versions should apear in 32 and 64bit versions.

stil hoping that MS will somehow merge the 32 and 64bit versions as else there would be alot of difrent Vista versions.

[the]

what i gathered from the artucle published in information week, of the 5 US versions, all will support 64bit except the barebaones one. have you not a 64bit machine, it'll simply run in 32 bit.

[Thorry]

940 - 1H03 (K8 server)
754 - 1H03 (K8 budget)
939 - 1H04 (K8 high performance)

AM2 - 1H06 (K8 level 2 budget)
F - 1H06 (K8 level 2 server)
M2 - 1H07 (K8L high performance)

History repeats itself often...

[largon]

I don't think K8L will need a new socket besides AM2.
It would make sense that the porc would work in AM2 with DDR2, and later in '07 (?) when DDR3 becomes available the same proc can be plugged in a fully AM2-compatible socket (named M2/AM3/whatever) on mobos with support for DDR3.

AM2 should last until Greyhound (desktop QC) = sometime in '08.
My :2cents:.

Hmm...
-- one of the strong attributes of our roadmap, both in 2006 and 2007, is socket compatibility.
- Henri Richards, AMD Executive VP and Chief Officer for Marketing and Sales (http://66.102.9.104/search?q=cache:ZGOSX9lJXmkJ:www.digitimes.com/news/a20060314PR200.html+&hl=en&ct=clnk&cd=2&client=opera)I wonder what Richards means by saying "both in 2006 and 2007"?

This reminds me of this post (http://www.xtremesystems.org/forums/showpost.php?p=1466885&postcount=48):

The reason I am bringing this is up is because I heard from people with contacts at AMD that the engineers designed the IMC of AM2 to withstand HTT speeds in excess of 500.Why on earth would AMD want to design the rev.F IMC to "support" freqs much higher than needed for DDR2?
This on the other hand reminds me of a certain debate some months ago...

[VulgarHandle]

well, isn't there a rumor goin about being able to change vcore for each core seperately? can am2 socket allow this? will this be a reason for the new socket?

also, wouldn't 65nm mean less power, and less power mean less need for pins....(although i'd hope they'd keep em for better regulation, so this may be mute, if not altogether ignorant)

i'm just tossin out q's..

[nn_step]

well, isn't there a rumor goin about being able to change vcore for each core seperately? can am2 socket allow this? will this be a reason for the new socket?

also, wouldn't 65nm mean less power, and less power mean less need for pins....(although i'd hope they'd keep em for better regulation, so this may be mute, if not altogether ignorant)

i'm just tossin out q's..
Yes each core can have a seperate voltage. and the reason for the new socket is for the electrical connection between the CPU and the RAM. And the Mechanical changes required for such connections.

[Thorry]

Socket AM2 will keep on existing just like 754 still does today, it's just not the mainstream socket by that time...

[VulgarHandle]

so then will it be M2(1207)?

[nn_step]

so then will it be M2(1207)?
and Socket F.. and Socket F is going to show up first and soon.. Its performance will make you either have to get a clean pair of pants or a new pair of shorts.. depending on which side you love...

[OmegaMerc]

and Socket F.. and Socket F is going to show up first and soon.. Its performance will make you either have to get a clean pair of pants or a new pair of shorts.. depending on which side you love...

Excuse me, but I use panties! :D :p:

[nn_step]

Excuse me, but I use panties! :D :p:
I can't pass up sig quoting that

[Dimitriman]

Arg, so socket M2 will have 1207 pins, cohexist with socket AM2, but only live for about 1 year before socket AM3 or whatever supports DD3 comes out??
sigh..

[nn_step]

Arg, so socket M2 will have 1207 pins, cohexist with socket AM2, but only live for about 1 year before socket AM3 or whatever supports DD3 comes out??
sigh..
NO Socket M2 WILL NOT HAVE 1207 Pins but It will have DDR3 support
(I have been saying this since christmas)
And please do not say M2 is AM2 :rolleyes: Cause that is a lie...

[Dimitriman]

NO Socket M2 WILL NOT HAVE 1207 Pins but It will have DDR3 support
(I have been saying this since christmas)
And please do not say M2 is AM2 :rolleyes: Cause that is a lie...

I believe you, but if thats true then my complaint still holds :(

[nn_step]

I believe you, but if thats true then my complaint still holds :(
Tis the Curse of the Overclocker.. upgrading and changing is inevitable...

[OmegaMerc]

I can't pass up sig quoting that

:idea:

:censored:

[nn_step]

:idea:

:censored:
:rofl:
I can't belive you just got that

[Pinnacle]

Excuse me, but I use panties! :D :p:

Merc, you freaky:D

[Thorry]

Please note:

The number of pins is almost absolutely determained by the amount of current (and indirectly the voltage) the CPU needs.

Socket F is a server CPU platform where high-power isn't such a problem and performance is top. Because it uses a lot of power a lot of pins are there and they have big surfaces (LGA).

Any desktop platform would not have such a great power usage (the noise and costs would be too high) I think something like 130 watt absolute max. The amount if pins would not have to go above 1000 for any kind of memory architecture.

Even if they integrate PCI express into the CPU they would prolly use a serial connection to the CPU to keep the pincount in check.

[Vincentvega18]

940 - 1H03 (K8 server)
754 - 1H03 (K8 budget)
939 - 1H04 (K8 high performance)

AM2 - 1H06 (K8 level 2 budget)
F - 1H06 (K8 level 2 server)
M2 - 1H07 (K8L high performance)

History repeats itself often...
I seriously like this reasoning.

[Pumbaa]

and the fact is that there is no big difference between 939 and 754 => M2 will suck?

[Thorry]

Well that's something nobody knows.

The difference between 754 and 939 is only dual channel, but it does make a difference for the product you can get.

There is no dual core 754 product, there is no 754 Opteron product, there is no 754 FX product.

If you compare the best 939 product to the best 754 product the difference between the two is most formidable, and not solely the difference between dual channel and no dual channel.

The DDR3 and 65nm process for M2 would mean it has a lot to offer... (also the other stuff we've seen).

But it is also logical to conclude that the best AM2 product and the most budget M2 product would be very close in performance (or even equal) just like is now the case with 754 and 939. This is because AMD wants to offer something in every price and performance range, not only the best of the best.

[Drunner611]

and the fact is that there is no big difference between 939 and 754 => M2 will suck?
Except, now we have DDR2.

[Bloody_Sorcerer]

"no big difference between 939 and 754"? how about these:
SSE3 on 939
90nm on 939 (widespread)
dualcore on 939
dualchanel on 939
better clock scaling on 939
to name a few.

[VulgarHandle]

well, there's 90nm on s754(3000+) and it has sse3 i believe(could be wrong, but it's rev.e, so i'm pretty sure it does)

[Thorry]

Yeah but check my post, the difference is huge because of the products released for the socket.

The only technical difference is dual channel versus single channel

[Vincentvega18]

I think the revolution between socket a and s754 was so huge that its hard to draw comparisons between that and 939->am2, however the process of bringing out the different product ranges in the same way would make sense for amd at this point.