savantu (savantu@email.ro) on 2/5/09 wrote:
---------------------------
>Vincent Diepeveen (diep@xs4all.nl) on 2/5/09 wrote:
>---------------------------
>
>>
>>Excuse me?
>>
>>When was the last time you ordered hardware?
>>What price would you want to pay for Tukwila versus Nehalem?
>>
>>>It is a problem however if money goes to Sparc/Power even if Nehalem
>>>has 2x the performance per socket and per thread.Sometimes that isn't enough and
>>>that's where the whole Itanium philosophy comes into play.
>>
>>Software gets upgraded in lots of companies. All kind of software from the past
>>already has been replaced long time ago. All that software already scales pretty
>>ok. It is NO problem to let single socket or dual socket machines do the job, as long as they got ECC.
>>
>>Very few softwareproducts can profit major league from 4 socket machines over 2 socket machines, pricewise seen.
>>
>>Basically we're speaking about AI engines such as my chessprogram where a 4 socket box is real cool to have.
>>
>>Note for 2009 even that isn't relevant much, as some total idiots decided single
>>handed to have a 8 core (logical cores) limit there.
>>
>>Yet there is a lot of other tournaments out there. Thing is, that is software that
>>doesn't run crucial services for companies now.
>>
>>Vincent
>>
>>
>
>By the same logic SPARC sales would be 0. Yet, according to a friend from Siemens,
>all serious SAP installations they do are on SPARC.
>Explain that.

Don't confuse legacy code that runs on a platform that total dominated highend thanks to Sun's dominant position.

Additionally there i a huge underestimation of how serious most big giants deal with their data.

Nowadays really a lot gets converted to seemingly weird formats like XML. Writing software for such formats is so easy though, that despite the huge overhead it has and slowdown in speed, which hardware has to make up for, companies really move forward real fast there to upgrade their software to levels that allows them to be flexible.

with a relative new and very bad supported platform called IA64 all this legacy SAP software is not there yet. It hasn't been made. Most itanium boxes i know of and where i ran at, they are all at organisations close to government, usually even university types.

All these organisations always try out hardware and pay bigtime for such trials. Yet odds they are going to buy tukwila is real tiny. If they do, it will be 1 machine and no more.

x64 is becoming a too dominant force simply.

Keep in mind also the huge recession that carves deep wounds. Budgets are a lot lower now. Price really dominates more than ever before. Instead of exhibistionistic luxury to order whatever hardware, they have to throw away a tad less money now.

Some years ago x64 / x86 was simply not accepted as a serious stable platform. That has changed however.

>Tukwilla doesn't need to beat x86 ; it needs to beat Power and SPARC to be as a

It definitely has to beat anything else, even if itis at just 1 terrain.

Licensing cost of middleware is serious business. Especially US companies i phoned asked amounts that were far above my yearsalary (right now anything is as mine is 0 currently as i'm without income). It is one of the few fields where being the best really matters.

Most burocrats, and i'm speaking about every nation now, are so surrounded by mediocrity and living a mediocre life that is so miserable, that the only thing that really turns them on is working with winners.

An Israeli, mr plausible deniability in own person, he once told me: "Winner is always right".

This is why in society winning is so important. These burocrats follow winners total blindfolded like sheep; it is very scary sometimes, as all disadvantages get washed away.

Intel positioned itanium platform as a big winner at the time. No matter its power consumption, no matter the huge amount of money a gflop, it was the FASTEST on the benchmarks and intel also was so clever to spread a lot of those machines long before release.

that means that every organisation and tiny company with some sort of software product that easily cross compiled for it, could already do some sneaky testruns here or there or at some spot on the planet.

So in all secrecy everyone already knew the huge ipc and big dang it would deliver. Then when it finally launched, itanium2 was really slow, simply because it was clocked at 1.5ghz and opteron was clocked 2.2Ghz and both had similar IPC's.

Fooling technocrats is real easy in this manner (giving somehow access at NDA'ed platforms and pray your competitor isn't doing it). That happened with itanium. That's how it was marketed.

Tukwila can't cause that same effect simply.

A 4 core chip that is going to compete against 8 core beckton? Now we didn't even take into account IBM nor AMD, which undoubtfully also are moving fast and offer great solutions.

>real cash cow for Intel.They're half way through this , with Itanium sales having exceeded SPARC sales last year.
>The obvious problem is that HP benefits the most from this and Intel should take
>steps to improve the situation of other IPF vendors , mainly Fujitsu and NEC.
>Of course, the golden goose would be IBM. :)

Cash cow?
Did you mean mad cow disease?

Vincent