Hacker News Reader

Yes, as an HPC admin, let me foretell how it will go here at our research group.

Me: Oh look this is fascinating! We have a worthy competitor to intel processors and we can do our work way faster. Let's get this!

Boss: Ok, tell me why one of this costs more than all of our other servers put together?

Me: But, it is so much better, it does blah, blah and more blah.

Boss: Do we really need it? What would we save by getting this instead of the commodity servers we normally buy. May be we should look into making our code work better with GPUs instead?

Me: ..........

Don't get me wrong. I think this really is fascinating. Just for the HFT traders and the top 1% buyers of exotic hardware. Not to an average HPC shop like ours, until the economy of scale and premium catches up to value.

>> Not to an average HPC shop like ours...

And there lies the problem. HPC used to be a tiny market with huge budgets where expensive hardware was the norm. It's just not that big a deal any more. It's commodity hardware now with lots of cores and GPUs running OSS. They won't make it trying to charge a premium just because it's Power9.

I wouldn't call Cray computers with x86 Xeon Phis commodity hardware though.

Cray CPUs are commodity tho'. The secret sauce is Aries, the interconnect. I ported code to the XC30 in a previous job.

I agree with your points [2][3]. Why is this so hard for vendors to understand?

It is an old idea called "mind share" [1] or in this case technical-mind-share, a vendor should want the largest possible community of technical people interacting with their technology and attempting to use it in as many niches as possible.

I'm not sure it even has to be uniformly cheap, as long as there is a low-cost entry point to hook people. If the price ratchets up from that low entry point but it is a compelling solution people will want it in their technical lives and get it funded somehow.

[1] https://en.wikipedia.org/wiki/Mind_share [2] I would point to RISC-V as another example of a newly popular machine architecture due to its initial low-cost. [3] noting the preferenced technology must meet the minimal acceptable threshold for all the usual dimensions.

As a corollary: see Microsoft, Adobe, etc. turning a blind eye to casual home users pirating their software. IIRC Microsoft even offered a free upgrade from pirated Windows 7 to a legit Windows 10 Home.

M68K and PowerPC architectures died out, and were available in consumer systems (Macintosh) and while not "dirt" cheap still in the ballpark of affordable. And you could find old cheap ones and run linux on them too.

The real reason Intel/AMD64 is popular is the same reason Windows is popular -- mindshare and economies of scale. IBM and the x86 platform won the business desktop in the 1980s and that has been propelling everything on that architecture ever since.

Dunno, not sure I buy it. From a software devel point of view it's "just" linux. The full ecosystem is there, clang/gcc, ruby, perl, python, javascript, apache, mysql, nginx, postgres, etc. About the only thing I can think of that might be different is if you are using cycle counters for profiling things, but it's rare to need to get to that level these days.

I think the real problem is everyone likes to target the top of the Intel systems that sell in very low quantities. Sure I'm sure HFT folks can make huge $$$ trading just a bit faster. But the bulk of Intel server sales are the low/mid range. Things like the E5-2620/E5-2630 that costs $300-$500 per CPU and has nice capable servers starting around $2k.

Sure $65k power system with 4 Voltas sounds great. But few will drop over $50k on a server without having a few cheaper servers in production for awhile. Beating the collective performance of 10 $5k servers is also going to be tough. Amusingly Intel fails that same test, rarely are the top chips the best price/performance unless the cheaper servers trigger needing to buy a new building.

Not to mention if IBM starts getting traction you can bet those $5k per CPU intel chips will start getting discounts rather quickly.

There’s a certain amount of laziness that prompts people to get bigger servers. I’ve been at a few companies where the solution to performance issues (as in, the thing that makes us money isn’t keeping up with customers and therefore we can’t grow our business) is to make the server bigger.

Switching to a distributed system is sometimes too far off the radar and costs precious engineering time. So the story is “just buy a bigger and faster computer, we’ll figure out how to shard this later.” The company drops more cash on IBM or whatever, and the engineers focus on features that close sales deals.

And then there are the larger companies which buy some expensive IBM servers and port their software in order to squeeze discounts out of Intel.

It's a difficult balance to get right. A $2k server as the GP describes costs less than a week of a fully loaded engineer. If those $2k buys you another X time (for any X >> week) of the application working, then it's not at all obvious that it's the lazy or otherwise wrong step to take.

(Lots of handwaving in the above, main takeaway is that it's the expected and rational path in a world where engineers are expensive and scarce and servers are cheap)

In this position.

Inherited a complete mess of a system (response time measured in minutes) running on an Xeon E3 w/8GB of RAM and 500GB of spinny rust.

It's going to take months if not years to unscrew some of the mess so I suggested to the boss the quickest way of improving it was to upgrade the server so we ordered a Xeon E5 dual processor system with 32GB of DDR4 and 1TB DC grade SSD.

Once I've ported the system from PHP5 to 7 on that and moved to a recent MySQL version we'll get a very nice perf improvement for basically free and then I can refactor the mess and at the end we'll have a better optimised system running on much better hardware.

Hardware is cheap, people's time isn't.

> Hardware is cheap, people's time isn't.

Sadly it's not always that simple.

People's time, salaries, falls into the base expenditure.

Hardware acquisition is a discretionary cost, which is usually squeezed particularly in a cost-centre like IT support. Or the existing hardware is under a multi-year lease from a vendor like Dell or HP and has to be paid-for whether it's obsolete or not.

Where I've worked ( big non-IT-orientated companies ) it's more typical to throw a couple of people onto a project to debug and refactor an application than buy better hardware. Or even worse, implement a new project on existing shared hardware.

Yes that can happen if people are already on staff and can be assigned to such an effort. But if the choice is spend $50K on new hardware, vs. hire three new devs to rewrite/optimize the software, the decision is easy.

Yep, there is only me and company doesn't want to hire a new dev so hardware is the cheapest way to speed things up now and it'll still be there when we are finished.

Salaries are taxed, hardware is deducted and amortized.

But the reason that intel is and has been cheap is the scale that their monopoly on the desktop has created. This position has given them a opportunity to develop fabs for node sizes that is out of reach for any competitor while keeping a competitive price point. At the moment it seems like Moores law is slowing down, and ARM on mobile has given scale to independent fabs to catch up with intel. So the situation may actually be a bit different this time.

It isn't about the cost of the cpu. Yes Intel eventually won on the market scale. But the real problem was, that you couldn't get a Sparc or PowerPc or whatever on a standard PC motherboard for a reasonable price. From the middle to end of the 90ies, lots of the Linux community would gladly have bought a non-x86 CPU/motherboard combo. The RISC-processors of that time still had an edge over Intel. Even later, with the G5 processor, they could have been competetive, if there had been a CPU/motherboard combo available.

Exactly this! Linus himself wrote that the only reason Linux exists is because a Sun workstation was too expensive.

I wonder if Moore's law was the real reason why all the RISC vendors failed while Intel succeeded.

Lets say IBM sells server hardware for 60K USD per machine versus Intel's 20k (I made the numbers up) and shows, through extensive benchmarking, that despite the higher initial cost, their machine's performance/dollar is better than Intel.

Then 2 years later Moore's law happens. Intel's new 20k offering now outperforms IBM's old machine on every metric. In hindsight, IBM didn't make sense financially.

With Moore's law slowing down, maybe now its the right time for IBM to make a move.

Intel chips are more RISC-like on the “inside” after the instructions are decoded, and in that sense RISC is successful. ARM is RISC and very successful. But most RISC and CISC vendors alike are failures by comparison to Intel and ARM, and the IBM Power processors are just not that performant per watt. You don’t even have to wait 2 years, they’re already not as good.

Part of this may be due to Intel’s lead in manufacturing technology, and have nothing to do with architecture.

(Historically, RISC was much more competitive. Talking about current state of affairs.)

ARM is RISC and very successful.

Actually, ARM is more CISC-y than MIPS, POWER, SPARC, etc. and recent cores break more complex instructions into uops just like x86.

If you consider uops RISC, then just about every architecture is "RISC-like on the inside". RISC is about a simple and restrictive user-facing ISA, not the internal microarchitecture.

In the long term, CISC makes a lot of sense --- it saves on fetch bandwidth and instruction cache to essentially have the CPU "decompress" complex instructions to execute internally, and the core is many times faster than memory.

IMHO "pure RISC" was an academic exercise, and the only reason why early CISCs were easily beaten was because they were sequential/in-order, and memory bandwidth wasn't a bottleneck at the time. With the growing core speeds, memory becoming a bottleneck, and invention of parallel uop decoding/execution, CISC could do more per clock and with less instructions. You can see this trend here:

https://en.wikipedia.org/wiki/Million_instructions_per_secon...

The best ARM core on that list can do 3.5 DMIPS/MHz, and the best MIPS at 2.3 DMIPS/MHz, while the best x86 core is at >10 DMIPS/MHz.

Have a look at slide 43 here:

https://riscv.org/wp-content/uploads/2016/01/Wed1345-RISCV-W...

This is from a couple years ago. Even more important (I can't find it right now) is that risc-v compiles code to fewer bytes than x86 and fewer micro-ops as well. With the smaller/simpler instruction set it also requires less area and power, so it may be just a matter of time now.

I guess the "even more important" thing you're thinking of is https://arxiv.org/abs/1607.02318 ?

So with RISC-V GC + macro fusion of a few common idioms you get slightly less micro-ops than x86-64, armv7, or armv8.

That being said, for high-performance cores such as Skylake, POWER9, or the latest aarch64 server cores the ISA probably doesn't matter that much.

Yes, that what I meant. But the small number of instructions on risc-v means less hardware to implement it. If it wins on code size and micro-ops but has far fewer instructions to deal with, it should result in smaller circuitry to implement. So far it has been implemented with less area and power consumption than ARM cores of similar performance and there doesn't seem to be a reason it can't scale to bigger/faster cores the same way x86 has.

I think the usual argument is that the decoder is a small fraction of the total transistor count in a high-performance core.

For microcontroller class HW, the x86 ISA might be a crippling disadvantage compared to RISC-V. For a high-performance core, which AFAIK is something like 20-30 million transistors, not so much. The bigger the core, the smaller the advantage of the ISA.

But yes, I don't there is any reason why RISC-V couldn't be used to create a high-performance core competitive with the x86, POWER, ARM of the day. It's just a hugely expensive affair.

While I agree that RISC and CISC is not much of a distinction in 2017 it's worth noting that that Cortex A15 is very far from the state of the art for ARM chips. The Cortex A75 would be about twice as fast, and Apple makes an ARM architecture that beats Intel's best (per MHz) in certain other single thread artificial benchmarks.

The big advantage that RISC had back when it came out was that you could fit an entire RISC core on a single piece of silicon, something that isn't really applicable in the modern day.

Right now the big advantage of RISC in the high end is ease of parallel decode thanks to instructions always starting on 32 bit boundaries. Which is maybe a 5% power advantage at most. 64-bit x86 has lots of historical baggage and the people doing 64-bit ARM were clever so they're actually about equal in instruction density. It was an advantage to have fewer instructions back in the heyday of RISC but it isn't any longer and both ARM and x86 have tons of different instructions.

When you move down from high power OoO cores to in order ones then having 32 registers instead of 16 becomes more of an advantage and the benefits of simpler decode become more significant too.

Agreed. Moore's law doesn't happen on its own and Intel has been leading the way in fabrication technology.

As for ARM, I think their success has more to do with their business model (licensing instead of fabricating) than with their architecture.

ARM is the future. Example Apple's A11. For single thread ((which is what's most important on 90% of today's use cases) side note can this stop being true by 2020?) can compete with single threaded processes.

Apple's A11 (single thread) 4217 (multi) 10164 https://browser.geekbench.com/ios-benchmarks

Intel i7 8700k (single thread) 6089 (multi) 26654 https://browser.geekbench.com/processor-benchmarks

It appears that we are near the end of physically making smaller nano meters and we are back to the co-processors (Like the old Amiga days) Co-Processors and some kind of Assembly (A Lisp that complies to an ARM Assembly to much for one to ask for???) are going to be the way we improve speed in the future.

Coincidentally (as no doubt you are aware) yesterday Microsoft released Windows 10 for ARM devices with built-in emulation, and a couple of hardware partners launched suitable systems based on SnapDragon 835 SoC. The future is arriving very fast.

Softbank's purchase of ARM is looking like a genius move even more now.

Yeah, but beyond that, I'm still of the generation (born 1981) that grew up hearing about amazing IBM RISC RS/6000 workstations, and so I'm still inexplicably drawn to any mention of this kind of technology as a “shiny thing”. I know it's mundane and broadly available in the lowest common-denominator hardware but it still attracts me like a moth to a flame. I know I'm a sucker.

ARM is the past, as pretty much every prominent chipmaker currently using ARM is in the RISC-V consortium today, and will be selling chips based on RISC-V tomorrow.

For the past year and a half, there has been no hardware for RISC-V. Where can one buy a cheap RISC-V desktop or a 19” rack mountable 1U server?

Furthermore, RISC V is one of the most awful designs I have ever seen in a processor internally. The assembler mnemonics are idiotic (compare and contrast with MC68000), the assembler is backwards (intel syntax of dst, src instead of src, dst)... it’s just an awful, non-orthognal, non-intuitive design. It’s an imaginary processor for imaginary hardware.

You can definitely get cheap MIPS machines, I have a MIPS board in my robot that only costs $5 in bulk.

Do you have a link to which MIPS machine it is?

Not the OP, but here are a few MIPS boards I'm familiar with in the under-$100 range.

Microchip makes a couple low-end MIPS boards:

http://www.microchipdirect.com/product/search/all/DM320103

http://www.microchipdirect.com/product/search/all/DM320004-2

When you outgrow those, you can move up to the Creator Ci20 from Imagination: https://www.mouser.com/search/ProductDetail.aspx?R=0virtualk...

> Zilog 80 baggage

8080, please. The 8086 was designed to be assembly-level source compatible with 8080's so porting software (here's to you, Gary) from CP/M would be easy.

> 8080, please. The 8086 was designed to be assembly-level source compatible with 8080's so porting software (here's to you, Gary) from CP/M would be easy.

1. The 8086 was not designed to be assembly-level source compatible, but nevertheless intended that assembly code from 8080 could easily be ported to 8086 (mostly find & replace)

2. Zilog developed a different assembly language for the Z80 than Intel used for the 8080 (for copyright reasons, I think - though nearly everybody would agree that Zilog's assembly syntax is better; compare for yourself: http://nemesis.lonestar.org/computers/tandy/software/apps/m4... ). The Z80 assembly language was a strong inspiration for the x86 (8086) assembly language (Intel syntax).

3. The Z80 introduced two index registers (IX, IY) over the 8080. A strange coincidence that Intel also did introduce two index registers for the 8086 (si, di).

4. Do the INI/INIR/IND/INDR/OUTI/OUTIR/OUTD/OUTDR that were introduced by Zilog for the Z80 strongly remind you of the INS/REP INS/OUTS/REP OUTS (just consider that the 8086 uses the direction flag) instructions of the 8086?

etc.

TLDR: Intel took more "inspiration" from the Z80 than is generally acknowledged.

The 8080 and 8086 designs draw directly from Z80. You can see it in the assembler mnemonics and how the operands are handled. And segmented memory access, oh my!

The 8080 predates the Z-80 by about two years.

He likely meant the 8086 and 8088.

That's possible, but, still, even though I find the Z-80 a rather kludgy and inefficient thing (compared to the simplicity of the 6502 and the 6502 machines we had at the time), if felt much neater than the 8086 family.

Actually it is possible to get cheap MIPS stuff, not only in anonymous Android devices but also (m)atx boards Lemote produce for the CCP (for obvious reasons).

Desktop is what buys mindshare. A Raspberry Pi is an embedded device, a desktop and a very entry level server all in one and exactly that is the appeal. Where can one buy a dirt cheap, modern MIPS 1U server or a desktop today?

Can they be purchased in US or Europe?

Sure: https://rwmj.wordpress.com/2015/04/29/mips-creator-32-bit-du...

They have abysmal performance however. An RPi 3 (or just about any 64 bit ARM board) will perform better and cost less.

> why does intel architecture with all of its Zilog 80 baggage dominate

The Z80 was a clone of the 8080, not the other way around.

You do understand that this is server not PC tech? The mainframes and minicomputer that service hundreds of users at once? You remember the old client server model? IBM System/370 IBM AS/400 etc before they became System Z etc?

But the machine shown here is not a mainframe nor a AS/400 or successor. It does not run AIX, OS/400 (pardon me: IBM i) or any of their mainframe OSes. It runs Linux.

And while I am sure it is a sweet machine that offers great performance and reliability, it is rather expensive. If you want to - especially if you mostly care about the GPU part anyway - you can get an Intel or AMD based machine for a fraction of the price.

IBM still stells the large POWER servers that run AIX or IBM i (in addition to Linux), and they still make their zSeries mainframes. But at least the latter are even more expensive.

All of these are really great machines; IBM has decades of experience building computers for some fairly demanding customers, and from what I know about these machines, it shows. But do not confuse the machine described in the article with IBM's top-shelf machines.

As someone who literally grew up on HPC (sgi Origin 2000 and 3800’s to be precise), yes, I completely understand.