Penryn and Nehalem

[macca22au]

Intel speaks of their tick-tock philosophy.

Every two or so years a big leap forward, the tock, and in the intervening years, the small improvements, the tick.

According to their site the next tick is penryn 45nm architecture, and the next tock is Nehalem as they call it, and they say is due in 2008.

However I can't establish from their site when the tick is due.  When will penryn processors be on the market?

At the moment they are marketing the 6850 Ultra, which I take to be the last iteration of the present 6700 series.

Can we expect the next tick, penryn, by Christmas this year?  Given NickN's fantastic sticky at the top of this board, I will take the next step when it is RTM.  I know the performance gain will probably still be incremental only, but given the nature of FSX it means more of the games potential will be revealed.

And late in 2008 in my last year of full time work, I hope that I can buy the first of the desktop Nehalems - and inescapably, upgrade my chipset, video card and all.  

I would be grateful if Intel could answer this request, or someone in the know.

[NickN]

Penryn has been out in dev for over six months. The first server based and high end desktop release will be November 12th however you may not see in on the shelves for consumer systems till Jan. That one is up in the air but if Intel is smart they will get that puppy on the shelves before xmas.

Nehalem just hit the dev scene at the beginning of September and will be out in 2008. When exactly is up to Intel and how far they need to go in making sure everything is peachy-keen. It could be ase early as Jan or as late as May.


The difference between the two is the fact that Nehalem will be the first Intel slug that matches AMD's 64 multi-core design. The processor is major architecture redesign of Intel's multicore. It will allow for configuration in such areas as core usage, cache, memory, threading and I/O. AMD's native multicore allows for such configuring of individual cores, memory and threads, but Intel's does not.

A major advantage to Intels 45nm shrink is in getting rid of the nitrided silicon dioxide gate insulator and move to a High "K" metal gate which removes the problem of power loss due to gate leakage which is associated with the SiO2 gate as dies are reduced. http://prola.aps.org/abstract/PRB/v49/i15/p10278_1

Nehalem will also bring the 8 core Intel slug to market.

In 2009, Intel will introduce the 32nm processor.




as a side note, I just happened to bump into one of the big-boys at MS who oversees the Server System dev teams in VISTA

Funny thing about that conversation... they know VISTA is a dog and they know why, which comes down to corporate having their nose so far up Intel's arse they were required to develop for IA-64 which with Nehalem becomes obsolete.

The bottom line is, the reason FSX and VISTA suck so well together is because of Intel shining MS into developing for IA-64 so they had an edge on keeping AMD's processor architecture restricted by the OS and leveling the playing field. Right now VISTA sucks the life out of everyones CPU effectively keeping AMD's old and newer processors in check with Intels and with the introduction of Conroe and other past year releases it gives Intel a slight edge over AMD until Intel can finish thier 'copy-cat' of AMD's 5 year old approach and put it on the market. VISTA essentially makes AMD's Barc and newer design run just as bad so Intel worked MS to do their dirty work.

get it?

It was a dirty trick because if MS had developed for x64 and the newer Intel/AMD processors, VISTA would not be as much of a dog as it is.


I came away from that conversation feeling as if there was going to be a major shift in VISTA by the end of next year, which is par because it was not until SP2 that WindowsXP began to shine.

[Brett_Henderson]

About 5 years ago.. I naively hoped that by now, LINUX would have found some software partners to be bringing us MSFS quality "games"...

A lean/mean LINUX istallation on today's computers would be scary fast (and efficient)..

(I was a XENIX Sys-Adm back in the 1980s)

[NickN]

My goodness, you were in the mix during the good-old-days.

[Brett_Henderson]

It's been nearly 20 years, so my memory isn't crystal clear..  :P   I remember the big fight circa 1986 over small office business. Networking (I think that's right about the time that the term "LAN" came to be), was the realm of Gurus and small businesses were learning that "computerization" wasn't all it was cracked up to be. If you didn't have a Guru working for you, you spent all the time (and then some) "saved" by computerizing, just keeping the network working.

UNIX was all the rage for multi-user .. and XENIX was its red-headed step-child. Microsoft's grab at the small "multi-user" market was OS2 (I think). The company I worked for was kind of pioneering, in that we turned away from DOS/MS/OS2 work-stations and went with pure terminals (VT-100s if I recall) all tied to an Altos server. I can't remember the version of XENIX that we used, but I remeber that ALL of the applications (AR/AP/PR/IC) were written in COBOL. The company who sold us the apps went under, and with a little arm-twisting, I got them to give me all the source code. Each terminal was connected via massive parallel cables and supported local printing (keeping that working aged me).. and the more remote terminals had these incredibly ornery modems which needed double twisted pair wiring (I'm laughing as I type this from a wireless laptop). The terminal nearest the main shop had to be sut down, and phyiscally disconnected from the server when they arc-welded... seems those twisted pairs gobble up the electro-magnetic cluster-foxtrot from the welders and force-fed it to the server .. LOL..

All in all.. it was a pretty amazing setup...  And I'll always remember how some of my other computing (Sys-Adms tended to find each other in those days) buddies would brag about their company's OS2 setups, running on IBM 80286 ATs.. and the one guy whose company would NOT give up their IBM S-36..  

OK.. hijacking done..  that just brought back memories...  I don't think all the big players have ever been on the same page with the end-user's best interest in mind...

[NickN]

Being 1985 XENIX was the answer to running affordable Unix V on a 286 and in 87 after SCO grabbed it from MS they added 386 support. Before that was an animal you guys really had to dive into in order to get network function out of it. Telnet or adding shared services was a no-go even after as I recall. The 87 update introduced TCP/IP but even with that, networking was still a grunt nightmare.

ATT had no clue what they had. After acquiring it, Gates tried to sell XENIX to IBM but they were afraid of what the ATT breakup may do the potential market.

I think the part that made Gates such an innovator was his ability to see how something like that could be used and how to market it the right direction. MS ran XENIX on their office systems right up into the late 80