Wednesday, January 31, 2007
Intel delays 45nm processors to 2008 !
Intel delays 45nm processors to 2008 !
Intel will position its 65nm-based Dual-Core Xeon 5100 sequence (Woodcrest) and Quad-Core Xeon 5300 sequence (Clovertown) series processors as the main force in the 2-way server market in 2007, and will not begin introducing its 45nm-based quad-core (Harpertown) and dual-core (Wolfdale) processors until 2008, according to sources familiar with Intel's plans.
Congrats. That day was however, quite enough for Barnum style marketing world around. Dual die monster at 2,6Ghz for only $900 in 1Q2008. Intel's obvious 45nm AMD killer. At unbeliavable, speedy 1066Mhz FSBus. :)
Don't expect that kind delay from yesterday shown (again?) AMD's quad core Barcelona below. Barcelona Opterons even now start to make rumbles. Thus, the actual battle is not in technology edge , but in economy of chip production.
However, do expect 45nm introduction at IBM in 1Q2008. Thus only 6 months advantage for Intel compared to AMD.
CHEERS. Anyway, I suggest Intel customers heavy use of this gadget.
Honestly, Intel 45nm efforts remind me on US efforts in Iraq, since war announcement in 2003. Win against nasty Iraqi terrorists is just around the corner, however we still need more troops and time.
So, victorious parade will be delayed a bit once again. And ...
Situation has been the same with Intel and AMD's Opteron announcements since 2003.
Saturday, January 27, 2007
Intel's new 45nm fraud
It is still fresh in our minds Intel's OFFICIAL fraud on 10-20 Ghz clocked processors, just around the corner, based on High-K and metal gates. This time, 10Ghz target is somehow dropped, but the other part of that fraud number one from past millenia remained.
"The implementation of high-k and metal materials marks the biggest change in transistor technology since the introduction of polysilicon gate transistors in the late 1960s," says Intel co-founder Gordon Moore.
Right, but what about the associated problems? 45nm itself alone is enough complex geometry to introduce sufficient number of production gremlins, even without High-K metal combination applied.
On the other side, Intel's some 40 years established bulk silicon production process is leaky enough to make Intel's standard SiO2 process ridiculous compared with AMD's SOI standards.
Intel did yet another fraud with 10Ghz. It REDUCED clock from former 3,8Ghz P4 to reduce power consumption and leakage in its dual cores. So, Intel masquaraded its max clock problem with NEW PROCESSOR ARCHITECTURE, that effectively replaced improvements that should be expected from its new 65nm process alone.
Though, 51 element is used in its production.
Now, Intel's marketing gun is empty and hard risky choice is ahead: To introduce radical new winning High-K metal process, or to suck AMD again in the next 2 years.
"It is claimed to increase transistor switching speed by 20%"
So, after "really big breakthrough" we will potentially get in the end shocking Intel's MAX 3,6 Ghz dual cores in 2009 !. In 45nm High-K.
"With 400M transistors"!
However, it is not stressed that Intel's 45nm core ITSELF has not necessarily more that 4M transistors !Why? Because the rest of chipbelongs to UNNECESSARY big cache, due to high IPC architecture, and and and ... LOW High-K process reproducibility. Consequently, that means in practice that new celebrated Intel's 45nm will be cheap and frequent like hen's teeth.
Intel first used the combination of metal gates and high-k dielectrics on the test SRAM (static RAM) chips it built using its 45-nanometer technology, Bohr said. The dielectric must be built using atomic-layer deposition, he said, meaning that a machine must deposit the dielectric just one atomic layer--the width of a single atom--at a time.
AMD will still remain cheaper, better and more competitive.
So in the end, where is essentially that Intel's breaktrough?
Only yet another Intel's big lie.
And why 45nm fab is displaced to Kiryat Gat, Israel?
No one will be able to sniff the truth.
And the truth is against the Intel: AMD is smarter choice by design, not only in 2H2007.
AMD processor has on average 2 times less transistors than Intel's. QUite enough for the same productivity with older geometry, compared to Intel. So, what new technology means for AMD? Nothing. They only have to be compeatable with their older SOI production geometry (compared to Intel's bulk up to now). And that is just what they are doing now:
Wednesday, January 24, 2007
AMD Barcelona dwarfs Intel's quad core performance
Really. Intel's two dies are mainly mutually blocked due to weak inter core bus. But, look at the Barcelona internal bus, at the center of chip. Obviously, there is something called system throughput, AMD insisit on its future multicores.
I pick the best, so there is no room for loyalty in high technology. To be precise, the marketing materials are predicting a 70% TPC? increase on a per-processor basis compared to Intel's essentially dual cores.
Ok, let say sustained 35% advantage might be quite enough. :)
Tuesday, January 23, 2007
Intel hires 1500 Jews, fires 6000 Americans and 6000 others
But, Intel will this year complete its plan to fire 10500 workers worldwide. If we now add 1500 Jews to be hired in Israel, that is actually at least 12 000 workers fired worldwide. However, 50% of Intel's workforce is out of US, thus it is substantiated guess that half of 12000, some 6000 Americans should be fired.
So, here is the final result of Intel's workforce reduction: 6000 Americans out, 1500 Jews in.
However, Intel's main $11B export from Israel is just to USA.
Let's compare it with AMD workforce politics. AMD hires in its New York fab, some 600 -1200 Americans.0 employed in Israel.
But, main marketing based impression about topics is that Intel is America's better admire company than AMD. Hah. Nothing more wrong than marketing and statistics.
Why shouldn't EU left Intel to sell their processors to country where it produces them and hires in situation when Intel's revenue dropped some 49% in financial year 2006?
Thursday, January 11, 2007
How IBM Power 6 will reach 4-5 Ghz clock ?
However, it is not a simple task. Thus, don't expect that from Intel soon. Why?
The inability to scale the tolerance of multiple electrical parameters along with their nominal value has contributed to a virtual crisis in the ability to improve performance and reduce power consumption in new processes. The continued infusion of new materials and structures provides an illusion of conventional scaling, but presents additional idiosyncrasies as well. Anticipation of these mechanisms and their influence on variability is critical. Circuit and architecture design innovation will enable the extension of CMOS technology beyond currently recognized limits. These lessons will be important in addressing the more profound challenges of novel emerging technologies.
But, expect it for sure from Power 7 in 2009.
Honey, I killed Mars Global Surveyor by ambigution of allocated memory objects
MGS - 2 Nov lost telemetry. Error messages on solar array gimbals received. Tried to image with Mars Express and MRO. No definitive indication that we found spacecraft. Have a tem looking at failure mode. We have some indication of what happened. Want to feed lessons learned into operation of other spacecraft at Mars. Close out activities will continue in May 2007.
MGS completed 34,202 orbits and returned 25,528 images during 10 years of operations at Mars. Final news form MGS showed that water may still be flowing on/near the surface of Mars.
We think that failure that a software load we sent up in June of last year was the cause. This software tried to synch up two flight processors. Two addresses were incorrect - two memory addresses were over written. As the geometry evolved. We drove the arrays against a hard stop and the spacecraft went into safe mode. The radiator for the battery pointed at the sun, the temperature went up, and battery failed. But, this should be treated as preliminary.
BAD comment: Shit, hack like this one has not been shown until now in object programming lessons. Is there little Osama in every of us? Anyway, we are eagerly waiting for home robots. They will spend a lot of time in idle, safe mode.
Tuesday, January 09, 2007
AMD K8L reference motherboard is alive !
AMD K8L reference mobo is alive. So, it will be commercially available in the next 6 months, and why not K8L itself immediately after?
Intersil works PWM for K8L processors, just like for Intel's latest dual corers.
ISL6322 supports both AMD and Intel. Why not forthcoming ISL 6323 used here?
Oh, and don't be confused. AM2G2 processor referenced here is a brand new K8L Star line processor.
Saturday, January 06, 2007
Computers obviously do cause global warming
Finally, a truth on global warming. Computer evaporates a bucket of water fastly.
However, if that bucket is a pool size, that fact is less evident for all of us. And if the bucket is say size of ocean, our problem is virtually non exiitant. But multiply the source of computer heath with a few hundred milions (so much computers work simulatneously each day). The heath problem then emerges again and associated fast evaporating too. Don't forget that only servers consume power production of some 40 nuke plants. And computing needs still grows exponentially. What will left in the end?
Boiling oceans. Oh, don't forget that water evaporate together with CO2 and methane, is the main greenhouse effect contributor. Thus, it is a sort of positive feedback amplifier for exponential rise of future global warming. That will hit the coldest places on our planet the most badly. They will start to melt fast too.
And bingo. Yet another global warming amplifier is turned on. The, boiling oceans are not far future from that point.