Bookmark and Share


Even though ARM Holdings, which architecture powers microprocessors inside the vast majority of mobile phones, is trying to enter the market of servers, the company does not seem to have plans to add a crucial feature for such systems: 64-bit capability.

"Some applications which run on service rely on 64 bit. Some rely on 40-bit addressing and that is built into Cortex-A15. Whilst Cortex-A15 is a 32-bit processor, it has the extended addressing. There are certainly server applications today for which 64-bit or lack of 64-bit is not a barrier. [...] We are not in the business of announcing our 64-bit products today. But it’s logical to suppose that at some stage in the future, ARM will extend its architecture in that direction," said Warren East, chief executive officer of ARM, during a conference call with financial analysts.

Server microprocessors of today are all 64-bit capable and so is the software. As a result, underestimation of importance of 64-bit capability seems to be more than short-sighted. Moreover, considering the fact that ARM does not seem to have immediate plans for a 64-bit architecture, it looks like the firm also does not have a priority to server market. In fact, this is not truly surprising, keeping in mind the performance of mobile chip market.

"It would certainly be helpful as and when we have those sorts of products. But right now, we are not ready to talk about those sorts of products and then large chunk of the server market is available with the product roadmap that we have," added Mr. East.

Tags: Cortex


Comments currently: 5
Discussion started: 02/06/11 07:02:59 PM
Latest comment: 02/08/12 03:40:03 AM
Expand all threads | Collapse all threads


ARM is eschewing the "double it" path AMD64 went; the ARM plan doesnt involve the same SIMD crunching plans that were being used for 3d math + video processing. ARM is very much about code size, and using smaller opcodes.

what ARM has announced already--this past summer-- is their plans for 40-bit (1TB) addressing-- an 8 bits addition from an all-too-small 32-bit 4GB address space to a seemingly reasonable 1TB.

this however seems a bit too small IMO. second gen SSD controllers will be pushing SATA III ports to their limits. sata ports are fated to disappear altogether as even faster simpler PCIe connected solid state drive controllers emerge. these drives will initially be in the 256GB/512GB domain, but with expanding capacity or multiple drives i expect ARM to exhaust it's physical address space really fast. this doesnt even go into the other part of ARM7a, virtualization, which will in ways both stress + alleviate memory address space pressure.

48-bit (6-byte) addressing would've been saner, would've given room for growth.

0 0 [Posted by: rektide  | Date: 02/06/11 07:02:59 PM]
- collapse thread

What does hard drive/SSD space have to do with memory addressing?
0 0 [Posted by: DavidC1  | Date: 02/07/11 03:29:11 AM]
I almost want to congrautulate you on wisely written text when i saw it's an real article on other site. Anyhow thanks for including source link.

Second thing is that 40-bit virtual address space shouldnt be too much of problem "for growth" in next few years and arm still dont attack onto server space. But this should be 40b real addressing not virtual as your quoted text says. And there's too much wrong approach feeling in that text no matter how well written. We should consider its eetimes and that they still praise 16bit priocessing for most of things while even 32b arm is way more efficient, and pricier ... but then 16b is overpriced and they forget to mention it in the last decade or so :wink:
0 0 [Posted by: OmegaHuman  | Date: 02/08/11 01:07:25 PM]

40 bit addressing refers to main memory addressing, not addressing disk storage. If it did, Win98, XP wouldn't work on HDDs larger than 4GB.

It makes sense not to go to 64 bit instructions, unless ARM wants to clearly split their product lines. 32 bits is more than enough for mobile applications for the short-mid term.
0 0 [Posted by: Zoomer  | Date: 02/09/11 07:13:40 AM]
- collapse thread

Yes, disk addressing is different to memory addressing, a file pointer doesn't have to fit in one machine register. But in cases where you are memory mapping large files, e.g. databases or other large data sets, it's convenient to use simple machine pointers.

Also, the ARM8 architecture actually does describe two very different sub-architectures for 32b and 64b implementations.
0 0 [Posted by: andrewmu  | Date: 02/08/12 03:40:03 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture