Does anyone think K8L will bring the crown back to AMD?

[Nocturnal]

I am a diehard AMD fanboy and I really do like their offerings. However, for some odd/strange reason, I don't think the K8L will be able to smash C2D like the way they have in the past.

Who else feels this way? This is coming from a seriously diehard AMD fan who has not had an Intel system for probably over seven years now.

 

[thilanliyan]

no.

 

[gOJDO]

IMO,
Only if K8L comes before Penryn, then it will bring back the crown of 1P & 2P PCs(desktop, workstation, server). But than, after few months, when Penryn arrives, the crown goes back to Intel
K8L will not bring the crown of mobile computers, whenever it comes.
For 4P+(be it with K8 or K8L), AMD will hold the crown until CSI arrives with Nehaleem in mid or late 2008.

 

[jpeyton]

Originally posted by: gOJDO
IMO,
Only if K8L comes before Penryn, then it will bring back the crown of 1P & 2P PCs(desktop, workstation, server). But than, after few months, when Penryn arrives, the crown goes back to Intel
K8L will not bring the crown of mobile computers, whenever it comes.
For 4P+(be it with K8 or K8L), AMD will hold the crown until CSI arrives with Nehaleem in mid or late 2008.

K8L is Q2/Q3 2007 product; Penryn is a Q1 2008 product.

 

[gOJDO]

We'll have to wait inorder to find out

 

[Stoneburner]

how can anybody possibly know? Maybe viditor with his secret inside knowledge....

 

[Tangerines]

It depends on how well it performs.

 

[justlnluck]

There are still fanboys around? I just buy whatever is best in my price range. Competition is good.

 

[Furen]

Originally posted by: justlnluck
There are still fanboys around? I just buy whatever is best in my price range. Competition is good.

There are always people that prefer a ceirtain brand over another, just like some people out there (me included) would buy an iPod rather than some other audio players that are vastly superior and cheaper. Just because you are rather pragmatic in your purchasing doesn't mean that everyone else is, or should.

To OP: From the little we know, I'd say that at the very least AMD will catch up to Conroe on a clock-for-clock basis in INT and will completely destroy it in FP performance. Notice the "clock-for-clock" thing, if AMD doesn't lenghten the pipeline a few stages I think K8L will be hardpressed to keep up with higher-clocked Penryns. I'm thinking AMD might even need to clock the memory controller at lower clocks than the rest of the core, since I think that is what is holding the K8's clocks back.

 

[gOJDO]

The ODMC has no impact on the freqfency of the K8. It is just the process and the the architecture which is not designed to clock high. Also, the K8 is designed to work on 130nm and then was shrinked to 90nm and to 65nm. K8L from the bottom is designed for the new 65nm SOI process. The ODMC on the K8L will be be improved:
- Write Burst & DRAM prefetching performance improovements
- DRAM writes can be buffered in the memory controller before being opportunistically bursted into DRAM controler to improve DRAM interface efficiency
- Read prefetcher detects stride paterns and issues prefetch requests based on confidence level
- Channel Interleaving
The DT/WS derivates on sAM2+ & s1207+(4x4) will support DDR2-1067.

 

[Munky]

Why would the K8L not bring back the performance crown to AMD? Unless AMD makes some major blunder, I'd expect the next gen AMD cpu to beat the Conroe at least in some benches, and match it in others if it can't beat it in all cases. I know Conroe is the latest and greatest cpu, but it's not powered by magic dust, and AMD engineers aren't stupid.

 

[Kur]

I thought I had read somewhere that the K8L had a different core design so that the processor could contain different functions and they could customize it per the customers request? Like 1 processor has xytw functions but customer A wants xykm functions. I don't know if that makes any sense or if they did say that.

(xytw & xykm mean nothing just using them as variables ftw)

 

[aka1nas]

I expect it to at least achieve parity with current C2D chips and hopefully penryn on a clock-per clock basis. I've been looking at C2D mobos for friend's builds and getting disgusted with all the problem most of them have. At least with current AMD systems, there is less on the board for one of the mobo/chipset manufacturers to implement poorly.

 

[gOJDO]

Kur, all K8L CPUs have same architecture: K8L. The differences between various mobile, desktop and server CPUs are:
- the number of cores
- the availability of L3 cache.
- HTT type, the number and the speed of the HTT link(s)
- the type of the ODMC(buffered or unbuffered) and DRAM speed support

 

[SickBeast]

It will, but only for a few months until the empire strikes back.

 

[soydios]

Only time and benchmarks will tell.

 

[Viditor]

Originally posted by: Nocturnal
I am a diehard AMD fanboy and I really do like their offerings. However, for some odd/strange reason, I don't think the K8L will be able to smash C2D like the way they have in the past.

Who else feels this way? This is coming from a seriously diehard AMD fan who has not had an Intel system for probably over seven years now.

Furen (as usual) has it right...though I don't know how much Penryn will be able to clock past K10 yet. Certainly it will clock higher, but the question is will it clock high enough to take back the crown.
I suspect it will, but not by much...as I've said before, it is shaping up to be a real ping-pong match for Intel and AMD over the next 2 years.

 

[aka1nas]

At least we'll get cheap Quad-cores

 

[Duvie]

Hey lets start the whole speculation game again!!!! (sarcasm meter on high alert)

Lets just sit back and wait for some reviews of real silicon and leave the speculation game to the market analyst.


I hope AMD gets right back to the top. It will better for all of us....

 

[Zap]

Originally posted by: Duvie
Lets just sit back and wait for some reviews of real silicon and leave the speculation game to the market analyst.

:thumbsup:

Some things we could reasonably predict, such as the C2D E4000 series. In this case we roughly know what performance and overclockability is going to be, and roughly know pricing/availability based on Intel roadmaps.

Some things we could not reasonably predict. Sure, people knew Intel was coming out with something new, but until they started flaunting demo systems and some people with engineering samples saw some high clocks, nobody knew for sure that C2D was gonna kick ass and take names.

 

[SparkyJJO]

I don't know but we'll find out.

What I find interesting is how ever since C2D it seems that the prediction is AMD has no chance to keep up now. Why not? AMD did it once, and pushed intel to #2 performance-wise for 3 years. So intel took the crown again (took them long enough to ditch that P4!) and people act like AMD is a goner now. Hasn't even been a year with intel on top yet.

People were disappointed with AM2 in comparison with C2D but AMD never said it was going to be that much better performing than s939. It was just a step towards something bigger (K8L). So far I haven't seen AMD release anything hyped up to be something it wasn't, they didn't hype up AM2 or the new 65nm chip. I haven't seen AMD fail to live up to what they said yet at this point.

I think that while Intel obviously hates AMD with a passion they still don't want AMD to totally go away. If they forced AMD to go under (not likely its gonna happen, just saying IF) then intel would be in trouble for monopolistic practices. I'm sure that if it wasn't for that intel would have pulled some stupid stunt to force AMD out and then charge us that $500 for a P2

 

[BassBomb]

Who cares who holds the crown? As long as well priced well performing products exist it doesn't matter who shells the most FPS or eats more SuperPi

 

[StopSign]

I doubt K8L will win back the crown. The lead Conroe has right now is just too big to overcome. When K8 came out it only beat the P4 by a little margin in performance. However it was a vastly superior architecture due to efficiency. The performance lead K8 had on P4 was nowhere near the kind of lead Conroe has on K8. I think the best K8L will do is match Conroe in certain applications but still trail behind it in overall performance.

Originally posted by: munky
I know Conroe is the latest and greatest cpu, but it's not powered by magic dust, and AMD engineers aren't stupid.

But Conroe is powered by magic dust, and then some. The kind of performance lead it holds on K8 is absolutely ridiculous. It is also a tremendously better value because a $180 processor, when overclocked, absolutely blows the top AMD chips out the water. There is nothing AMD can do right now with the K8 architecture to overcome this. K8L will definitely close the gap but don't expect it to dominate Conroe and all of a sudden everyone hops back on the AMD bandwagon.

A lot of people have forgotten how AMD took the crown with K8. The performance was actually head to toe between Barton and Northwood for a long time. It was 2.4C vs XP 2500+. AMD then came out of nowhere and released the K8 which set new standards for efficiency and offered a healthy performance boost. It didn't blow NetBurst away like Conroe did to K8 in both performance and efficiency.

 

[AkumaX]

what's after K8L?

 

[Cogman]

I think it probably will, but I don't think they will hold it for very long. Intel has a lead and it looks like they are going to take every advantage they can with it. With the new 45 nm looking to be released on time, or possibly even early, I don't think amd can pull a big enough lead to counter act the benefits of the new 45 nm process Intel is almost finished designing.