New IOP XScale boasts RAID, dual port RAM support

By Chris Williams. Published: 22nd Mar 2005, 13:29:22 | Permalink | Printable

It's like family to us

Intel have recently unveiled the latest addition to their I/O Processor family, the 800Mhz IOP333. Billed as a storage processor, the chip includes RAID-6 acceleration, allowing it to manage reliable data storage on large disc arrays.

The XScale IOP333 is software compatible with the 600Mhz IOP321 CPU present in Castle's current Iyonix computers, and like the IOP332, the IOP333 doesn't have the bus performance issues that the IOP321 part is believed to suffer from. The new IOP333 includes a faster internal bus (333Mhz, shifting 2.7GB/s), dual ported DDR RAM support and more interrupt pins than the IOP332. Video memory is commonly dual ported RAM, as it enables data to be read and written effectively at the same time, and it's understood that MicroDigital are using dual ported memory in the Omega computer.

As a storage support processor, the IOP333 is part of Intel's roll out of their 64 bit Xeon enterprise range of hardware.

Links

Intel's I/O Processor family

Previous: Last call for pre-installed Iyonix Linux
Next: News in brief

Discussion

Viewing threaded comments | View comments unthreaded, listed by date | Skip to the end

Whilst the increased Mhz, improved bus and dual ported RAM surely could be of benefit to RISC OS systems, it sounds like RAID-6 acceleration will be of little relevance.

I wonder why Intel have chosen their ARM processors to become disc controllers. Have they been used for this sort of application in the past?

flypig on 22/3/05 2:14PM
[ Reply | Permalink | Report ]

flypig: ARMs, and ARM clones are very very frequently used in embedded fields. For unsurprising reasons, next to nobody uses them in desktops. IOP does stand for I/O Processor, after all - it's a microcontroller. This one just has a slightly specialist purpose, though.

What does PXA (the range od XScale often used in handhelds) stand for, anyway?

I'd be amused if one of the RAID cards based on this found its way into an Iyonix, though.

nunfetishist on 22/3/05 3:26PM
[ Reply | Permalink | Report ]

PXA = Personal eXchange Architecture (possibly).

jms on 22/3/05 3:58PM
[ Reply | Permalink | Report ]

It was my understanding that RAID wasn't an acceleration at all. It is, as the article says, deigned for 'large disc arrays'. RAID 6 in particular is suited to very large disc arrays - so much so it's inefficient with small arrays (i.e. one or two discs).

I can certainly see how this is an advantage for Xeon hardware, as they're primarly used for servers.

Otherwise, let's hope we see some of these new processors in a new Iyonix (fingers crossed). I suppose a disadvantage of using processors designed for the embedded market is the pin layouts keep changing, making otherwise cheap upgrades, like swapping a P4 2.8GHz for a P4 3.2GHz HT on a PC, very difficult.

Smiler on 22/3/05 4:55PM
[ Reply | Permalink | Report ]

I've seen a lot of SATA RAID cards featuring XScales.

simo on 22/3/05 4:57PM
[ Reply | Permalink | Report ]

No one's said RAID is an acceleration - although in fact, it can be depending upon the type.

The article said "RAID acceleration" - RAID can involve processing, which is what's being sped up.

mrchocky on 22/3/05 5:07PM
[ Reply | Permalink | Report ]

flypig:

Intel have been using RISC processors for these types of applications for years.

See what a search on "i960" brings up - I seem to remeber seeing intel RISC chips with this moniker years ago....

ROHC on 22/3/05 5:51PM
[ Reply | Permalink | Report ]

I was using Mylex RAID cards with i960's on around '99.

piemmm on 22/3/05 6:03PM
[ Reply | Permalink | Report ]

This is interesting. However, the performance boost over current Iyonix is, I suspect, too little to tempt Castle into an upgraded computer.

Raw compute goes up by 33% (600MHz to 800MHz) plus whatever gain the improved bus gives. The original StrongARM upgrade showed that RISC OS is actually very good at running from cache, so this is unlikely to speed basic OS and typical application performance very much. Hence, the overall boost is unlikely to reach 50%; 50% does not feel like much when you're sitting at the keyboard.

Anyone find a fallacy in my logic?

TonyStill on 22/3/05 7:08PM
[ Reply | Permalink | Report ]

Yes, please refer to previous articles were we've pointed out that the IOP331/332/333 will be much faster than the 321.

It's one of those things that people keep ignoring no matter how many times we state it. I wouldn't personally use the StrongARM comparison as more than a weak guideline.

mrchocky on 22/3/05 7:25PM
[ Reply | Permalink | Report ]

mrchocky: agreed

what will make this one faster is *not completely* the CPU speed its the faster memory & PCI support.

How fast the CPU for us is largely irrelevant.

What is attractive is that this chip is designed for data bandwidth.

We don't need large CPU overheads - the key to a proper modern RISC OS computer (with the usual caveats of the ADFS/CDFS limitations) is faster data in & faster data out.

Now who's going to twist Castle's arm(s)? ;-)

ROHC on 22/3/05 7:43PM
[ Reply | Permalink | Report ]

Didn't John Ballance (or someone else who should know) point out that the IOP331 would require a complete m/board redesign? Presumably the IOP333 chip is equally different physically to the Iyonix's IOP321? That said, I hope Castle /do/ have plans for an Iyonix2, if only to keep up with the VRPC clones.

George

bucksboy on 22/3/05 8:04PM
[ Reply | Permalink | Report ]

yup the new chips are different in pin out - so this would need a new m/board...

Now who's going to twist Castle's arm(s)? ;-)

ROHC on 22/3/05 8:35PM
[ Reply | Permalink | Report ]

Why would it need to be Castle? STD were looking at producing a home computer (A9home), maybe they will take advantage of this latest part and jump ahead of the game?

dgs

dgs on 22/3/05 10:41PM
[ Reply | Permalink | Report ]

Incidentally, if the Xeon is "enterprise", what's Itanium? :-)

dgs

dgs on 22/3/05 10:44PM
[ Reply | Permalink | Report ]

dgs: I think the A9 is going to be built around the ARM9 chip...No?

George

bucksboy on 22/3/05 10:52PM
[ Reply | Permalink | Report ]

bucksboy: "I think the A9 is going to be built around the ARM9 chip...No?"

Yes, no, don't know. Pick one!

dgs

dgs on 23/3/05 12:00AM
[ Reply | Permalink | Report ]

dgs: I'm not sure, but they're very very very quick at RSA. I know of a server room with bags of different architectures and CPU power (from HP-PA to SPARC, and MIPS to Alpha), and not only did the single Itanium in there beat them all at SSL speed, it was faster than all of the others put together. Perhaps it has an "RSA" instruction?

Other than that, they appear to be generally cack and expensive.

nunfetishist on 23/3/05 2:23PM
[ Reply | Permalink | Report ]

In reply to nunfetish: That is obviously why IBM have all but ditched Itanium in favour of the Xeons.

sa110 on 23/3/05 6:43PM
[ Reply | Permalink | Report ]

Please login before posting a comment. Use the form on the right to do so or create a free account.

+++ Message board +++

Read messages or start a new thread

Search the archives

Today's featured article

Online banking with RISC OS
We thought this would be of high interest

28 comments, latest by devine on 25/9/05 8:30AM. Published: 1 Sep 2005

Random article

UniPod speed tested
IDE and ethernet with a need for speed

28 comments, latest by micken on 17/8/04 3:53PM. Published: 2 Jun 2004

Create a new account
Forgot your password?

Useful links

News and media:
Iconbar • MyRISCOS • ArcSite • RISCOScode • ANS • C.S.A.Announce • Archive • Qercus • RiscWorld • Drag'n'Drop • GAG-News

Top developers:
RISCOS Ltd • RISC OS Open • MW Software • R-Comp • Advantage Six • VirtualAcorn

Dealers:
CJE Micros • APDL • Castle • a4 • X-Ample • Liquid Silicon • Webmonster

Usergroups:
WROCC • RONE • NKACC • IRUG • SASAUG • ROUGOL • RONWUG • MUG • WAUG • GAG • RISCOS.be

Useful:
RISCOS.org.uk • RISCOS.org • RISCOS.info • Filebase • Chris Why's Acorn/RISC OS collection • NetSurf

Non-RISC OS:
The Register • The Inquirer • Apple Insider • BBC News • Sky News • Google News • xkcd • diodesign

© 1999-2009 The Drobe Team. Some rights reserved, click here for more information
Powered by MiniDrobeCMS, based on J4U | Statistics
"I was actually referring to drobe itself as the 'young upstart' rather than the age of the contributors"
Page generated in 0.0726 seconds.