[ShrimpBrime]

Originally Posted by BenchZowner

What they said, and will supposedly be doing is not FOCUS into high performance desktop CPUs.
As in, they'll still release new chips, etc, but won't be the product of a "huge effort" in order to tie or surpass Intel.

In other words, expect BD v2, BD v3 ( with both being significantly slower than the new Intel CPUs that will come ).

They won't be as fast as people want. It's got less transisters per x86 core. Simple as that.

The 8170 released Q1 (supposedly) will be Octo with 3.9ghz clock and 4.5ghz Turbo and probably do 5ghz or better daily for extremists. I believe the new stepping will be C0 and should be a better all around cpu.

If AMD found a way for the actual module to act as a single core, it would be a fast cpu.

[BenchZowner]

There's no way to make a module ( the term is compute unit, there's no module in any microprocessor engineering book ) appear/act as a single core.

Also expecting to see an eight-core FX working operate fully stable for 24/7 usage over 5GHz on air cooling is more than a daydream, and even if they ( let me enhance the dream ) could make BD like 10% faster with a revision [ simply impossible, just saying ] and even make them have a decent average air cooled overclock at approx 5.5GHz, it'd still not be enough to match the 2600K.

[MaadDaawg]

The K6 was a hot processor

[ShrimpBrime]

Originally Posted by BenchZowner

There's no way to make a module ( the term is compute unit, there's no module in any microprocessor engineering book ) appear/act as a single core.

Also expecting to see an eight-core FX working operate fully stable for 24/7 usage over 5GHz on air cooling is more than a daydream, and even if they ( let me enhance the dream ) could make BD like 10% faster with a revision [ simply impossible, just saying ] and even make them have a decent average air cooled overclock at approx 5.5GHz, it'd still not be enough to match the 2600K.

10% is not that much. The last x86 (phenom II) architecture gained 10%. Why not this one?? C2 and C3. Big difference. Night and day actually.

__Cpu__Stp___stock___MaxTurbo
FX-8150 B2 3.6 GHz 4.2 GHz
FX-8170 C0 3.9 GHz 4.5 GHz

The above is only 60mhz shy of 10%.

I don't do the Intel vs AMD fyi. Already know i7 is faster.

[BenchZowner]

Phenom II is not just a process revision of Phenom I.

Clock per clock ( better performance at the same frequency ) is nowhere to be found in almost every single core revision ever.
Core i7 920 C0 vs D0... clock both at the same frequency, and show me a difference in performance
Core 2 Quad Q6600 B3 vs G0... clock both at the same frequency and show me a difference
...etc

[Gutterz]

Seems that their is 800 million less transistors than thought on Bulldozer.

source
http://www.tweaktown.com/news/21780/...ght/index.html

[BenchZowner]

Allow me to doubt their new number.
1.2B transistors is unbelievable, especially with those huge caches.
Chip must be around 1.5-1.7B in reality ( looks like AMD thinks numbers aren't important [ transistor count, pricing, performance figures ] )

[MaadDaawg]

: AMD :

[ShrimpBrime]

Originally Posted by BenchZowner

Phenom II is not just a process revision of Phenom I.

Clock per clock ( better performance at the same frequency ) is nowhere to be found in almost every single core revision ever.

Has the x86 core changed in forever? Nope.....

Trinity should be released soon (APU) with a presumed up to and exceeding 20% performance increase over current A8 APU's.

Aren't the i7's APU based chips? Not just cpu's but also have IGP right?

How will these Trinity APU's compare to i7's? Is it all strictly about cpu performance? Or is a cpu worthless with garbage IGP?

JW cause I know very very little about Intel. It's beyond my budget capabilities.

[ShrimpBrime]

Originally Posted by BenchZowner

( looks like AMD thinks numbers aren't important [ transistor count, pricing, performance figures ] )

Intel did this with NetBurst....