[Info-vax] "x86 has only a few years left in the market place" (was: Re: Free Pascal for VMS ?)

[Kerry Main]

> -----Original Message-----
> From: Info-vax <info-vax-bounces at rbnsn.com> On Behalf Of Stephen
> Hoffman via Info-vax
> Sent: May 12, 2018 12:27 PM
> To: info-vax at rbnsn.com
> Cc: Stephen Hoffman <seaohveh at hoffmanlabs.invalid>
> Subject: Re: [Info-vax] "x86 has only a few years left in the market place"
> (was: Re: Free Pascal for VMS ?)
> 
> On 2018-05-12 14:57:53 +0000, seasoned_geek said:
> 
> 
> > ...You are a limited sized team with a limited amount of money. The x86
> > has only a few years left in the market place with ARM being the new
> > bottom and aggressively consuming all of the low hanging fruit in its
> > upward journey, just like the x86 did.
> 
> Nobody likes x86-64.  Not even the folks at Intel, judging by the
> number of times they've tried to replace it.    But x86-64 is the
> volume desktop and server design at present, and it'll be around and
> available for the foreseeable future.   And it's what the vast majority
> of server customers and hosting providers have consolidated on,
> directly and via virtual machine guests.   Many didn't really like that
> Alpha and Itanium were different or different enough from the ProLiant
> or Dell or SuperMicro servers they were running.
> 
> The costs and support and software to adopt and adapt to a bespoke
> hardware platform is one more impediment on continued use and new
> adoption for any closed-source commercial operating system package.
> 
> x86-64 has a few years left in the market?  Nope.  The ~billion x86
> systems in present use and with uncountable numbers of applications
> and
> dependencies and users, and a morass of the current scale does not get
> ported in less than a human generation or two.  And there'll be
> stragglers through the rest of the century.  Which means servers and
> products will remain available.
> 
> As for options and alternatives?   The folks at VSI are pretty savvy.
> They looked around.   They were already looking around back before VSI
> became public in 2014, too.
> 
> As for POWER?   It's too expensive and too high-end and with no obvious
> path toward vastly larger production volumes in the next decade if
> ever.   It's not a replacement for x86-64.  It's also not at all clear
> that IBM will continue to pour buckets of money down the same
> bespoke
> processor effort.  That's a risky bet for a third-party software vendor
> such as VSI, at best.  IBM and POWER haven't looked like a particularly
> robust business over the past several years, either.   IBM did just
> recently have a pair of revenue-positive quarters, after their recent
> multiple-years-long revenue declines.  Whether those recent
> revenue-positive quarters were due to exchange rates or fundamental
> changes at IBM remains to be determined.  Consolidation and all...
> 
> As for Arm and maybe RISC-V, they're in a vastly better market position
> than is IBM and POWER.   But Arm doesn't build processors.   Other
> folks create the processors, some generic and some bespoke.  Arm
> servers were not in a tenable position when the OpenVMS port started
> back in 2014, and the folks making Arm servers and software just
> starting to have server configuration and toolchains and related.  And
> back in 2014, Arm servers were a whole lot less obvious, and important
> details such as AArch64 and SBSA — SBSA provides consistent processor
> and system designs for servers, a very important detail for
> cross-platform operating system software — and 64-bit addressing were
> just firming up back in 2014, and the open source tool chains
> themselves , when the key decisions were to be made.
> 
> Yeah, you routinely rage about open source, but no vendor can create
> and maintain completely bespoke and competitive software systems and
> remain economically viable.  Not even Microsoft.
> https://open.microsoft.com  DEC couldn't manage that back in the
> previous millennium, and VSI is a whole lot smaller than DEC in most
> any dimension measured.
> 
> VSI is well aware of Arm and indicated such back when that whole
> porting show got started in 2014.  As a potential future porting target.
> 
> Any suggestion of shifting to a different platform years into the port
> and right around first boot borders on a decision to kill VSI and
> OpenVMS.  Arm may well overrun the server market in the next
> decades,
> but it'll take a decade before Arm servers become widespread at best,
> and the x86-64 servers are going to continue to be available in vast
> numbers and myriad models for the foreseeable future.  And Intel will
> have more competition than they might prefer for their pricing.
> 
> OpenVMS on alternative platforms?   Sure.  Maybe.   Call back in five
> or ten years, once the OpenVMS x86-64 port is available and robust, and
> once Arm servers or RISC-V — or who knows, maybe even POWER —
>  become
> far more available and far more ubiquitous and acceptable server
> hardware platform alternative for a small software-only company.
> Adding additional supported platforms makes more work and more
> complexity for VSI and ISVs and end-user customers, and there needs to
> be a good reason for that cost and that porting effort and those
> hardware and software support costs.  And marketing both x86-64 and
> Arm
> or whatever else makes for more complex marketing for VSI, too.
> Replacing platforms?  Arm and POWER and RISC-V are not viable
> replacements for x86-64 servers.  Not yet.
> 
> x86-64 is what VSI is depending on, and there'll be increasing
> dependencies on open source and extending well beyond LLVM.  Where
> we
> are once the x86-64 port is settled?  We shall see.  I'd not be
> immediately looking for another port either, as there's a whole lot of
> work necessary to bring the OpenVMS platform forward and to broaden
> peripheral and hardware and software support, and a platform port — at
> the scale of VSI — freezes most of that very necessary work.
> 

All have their own markets, but longer range, If any of existing architectures are at risk, I would say it is ARM.

What Qualcomm’s Exit From Arm Server Chips Means - May 10, 2018:
< https://www.nextplatform.com/2018/05/10/what-qualcomms-exit-from-arm-server-chips-means/>

"Since the Avago-Broadcom deal went down, a lot has happened, and it has all affected the Arm server chip ecosystem in one way or another. AMD has silenced (but not officially canceled) its “K12” Arm chip project to focus on its Epyc X86 server chips, which frankly present a healthy alternative to Intel’s Xeons without requiring companies to change instruction sets and therefore have been their own kind of deflation of the potential for Arm server chips."

In terms of PowerX, it seems to be doing ok. Its market is not to replace all Intel based Server computing, but rather focuses on the higher end Server spectrum.

"IBM Rounds Out Power9 Systems For HPC, Analytics" - May 09, 2018:
<https://www.nextplatform.com/2018/05/09/ibm-rounds-out-power9-systems-for-hpc-analytics/> 

"A lot of things have happened in those 25 years, and one of them is that IBM turned around its Power chip and system business and eventually vanquished rivals Sun Microsystems and Hewlett Packard from the glass house. To be sure, the Unix server business is considerably smaller than it once was, but IBM has embraced Linux and has created a compelling set of processors, with lots of I/O that is not available on Intel Xeon or AMD Epyc processors, that give it a credible chance of stealing some market share from X86 platforms."

It should also be noted that Google has added Power9 to a number of its performance sensitive App environments.

" Google Confirms POWER9 Processor Data Center Deployment At OpenPOWER Summit 2018" - Mar 09, 2018
<https://www.forbes.com/sites/patrickmoorhead/2018/03/19/headed-into-its-fifth-year-openpower-has-momentum-into-the-power9-generation/#59cc6a7678a8>

Short version of url:
<https://bit.ly/2jRracF> 

Summary - Intel will continue to dominate the DC Server market, but Arm and PowerX will likely carve out speciality niches for themselves in specific areas.

Regards,

Kerry Main
Kerry dot main at starkgaming dot com