/tech/ - Technology★

>rs-232 port

I looked. Really hard. For any peripherals that still used an rs-232 port. I couldn't find any, other than rs-232 to USB2 converters. Why bother including the port on the back panel? Why not just a UART pinout if they wanted a hardware terminal? Also, it could have done with an m.2 pcie slot or two, to maximize that coreboot boot time.

I'm considering buying one, but I'm going to wait for a review or five of the board before i make a decision, because i'd have to drop around 4k-4.5k making a PC around it, considering e-atx cases and ddr4 ram prices.

Also, why the fuck did they give it 16 ddr4 ram slots? you can get 128g ddr4 ram sticks.

>>788938

For 2 TB RAM support, dumbass

ok I wrote that then checked the spec

>Supports up to 2TB total main system memory

Holy shit

wtf is Power9? Why would I want that? I might if most languages run/compile on it and the laptop has a mech keyboard

>>788797 (OP)

>Anyone who is interested in the Talos II?

I would be if I wouldn't have to take out a second mortgage to get one.

>>788797 (OP)

I'd buy one no question if I wasn't unemployed.

>>788944

are you stupid or something?

>>788995

just /tech/ my man

>>788944

>>>/g/

I managed to find a couple old keyboards that use rs-232 connectors. Other than that, nothing. I was hoping i could find an old terminal screen/keyboard combo, but nothing usable turned up. I found one or two old keyboards that used the proper port, but why bother? Why the hell did they include 2 rs-232 ports? It seems like an utterly deprecated standard to me.

>>788938

>rs-232 port

any industrial equipment made for past 30 years, hardware debug

>it could have done with an m.2 pcie slot or two

Like proprietary intel cut dick much? NVMe bringup firmware is non-free and needs UEFI, thus mitigating estonishing coreboot boot times.

will buy if its still available when i have enough money

My company is considering this for batch processing security camera footage. Were probably never going to see it.

Why does the mobo cost 2200 dollars?

Aren't these CPUs mainly for server applications? I thought we were waiting for CPUs based on the RISC-V instruction set to be manufactured for consumers to curb our reliance on x86.

>>789878

do you know how fucking long it will take for riscv to get to this level?

Power9 is open source and works now. You can compile any open source app for it.

This is true open source.

>>789876

Because it was custom made and manufacturing for this level of hardware is expensive.

>>789886

Haven't done that much research into it, sorry. Are there any initiatives for affordable consumer grade hardware components?

>>789888

Not for open source hardware. And keep in mind this is not just open source it is a powerful system. This is the first PC with PCIe 4.0. Personally I am going to buy a motherboard and wait for the 24 core power9 cpus to come out. Thats a beast-level open source system. Dream come true for /tech

>>789886

>>789876

Talos II is still priced similarly to a similarly-performing Xeon workstation. The first Talos failed, and it was more expensive.

>>789876

>>789878

The mobo costs 2200 dollars because of the developement and manufacturing cost, especially because of the presumably small batch size.

The cpus being made for servers makes little to no difference to its effectiveness at general comuputing. My bet is that IBM skimped on desktop features (like more than just a framebuffer for integrated graphics)

Other than graphics, server processors mostly end up doing similar work to desktop processors. And the talos 2 is priced roughly around similiar i9 and xeon dual-cpu setups, including cpus.

>>788938

>I looked. Really hard. For any peripherals that still used an rs-232 port.

How to out yourself as a pleb in only 13 words.

>Also, why the fuck did they give it 16 ddr4 ram slots? you can get 128g ddr4 ram sticks.

Because each CPU has 8 independent memory controllers, also even 2TB of RAM disappears quickly when you are running lots of memory intensive VMs.

>>789876

Dual socket enterprise-grade motherboard + small production run = expensive

>>788938

>doesn't work in a place that deals with RS-232

Go back to writing ruby apps for food.

>>789276

>2 rs-232 ports?

I see one RS-232 and one VGA.

>>789890

Yeah, I have a boner.

I just wish it hat more usb ports for all my pleb peripherals.

>>789876

Because they don't use the Intel razor+blades sales model of selling you a trashy $200 mobo and then charging thirteen thousand dollars for the CPU

>>789890

>This is the first PC with PCIe 4.0

Its also the first PC with CAPI, which is even better than PCIe 4.0.

My current job pay would mean saving up for 3-4 months just for the mobo. Still waiting on the EOMA68 I preordered as well.

>>791813

>That said, the price is a steal for the kind of performance you get

Yeah, the price is similar to a similar-performing Xeon-based workstation. It's much better than the failed original Talos.

Why is it so expensive? The type of people who use "workstations" generally need Windows and other proprietary software. Companies are unlikely to switch their servers to an untested platform.

They should have made it much less powerful and marketed it to hobbyists/freetards.

>>791914

>The type of people who use "workstations" generally need Windows and other proprietary software.

A lot of scientists and engineers use UNIX-like systems so they often use Linux. Also Linux is dominating professional 3D workstations (Weta digital, ILM, Pixar, DreamWorks), those people write their own software or use software that works natively on Linux (Nuke, Maya, DaVinci Resolve, Lightworks, ...). Windows workstations are mid-tier , since OS is closed source it can't be tweaked, high profile jobs/research is often done on some sort of UNIX. Market is certainly there the question is what offers better performance per $, so companies will buy that.

>>788797 (OP)

V A P O R

A

P

O-

R

W A R E

A

R

E

>>791925

When you can compile your own software library for it from the vast selection of FOSS software you won't ever have to worry about vaporware. Unless you are referring to it becoming hardware that is no longer produced in the future?

Bump for interest.

>>791914

>Why is it so expensive?

We could ask the same of Windows, which is worthless

What makes you think Power isn't backdoored just like x86?

>>791914

>Why is it so expensive?

You haven't seen the first prototype.

>Companies are unlikely to switch their servers to an untested platform.

Because you could audit the X86 platform ?

That's something that I wasn't aware of :^)

I understand what you meant but in reality most techs (except oldfags from the 80/90s) don't know what's happening besides "it just works".

>They should have made it much less powerful and marketed it to hobbyists/freetards.

Do you know in what environment freetards work ?

I mean those who have a job ?

They work in datacenters, for ISPs or scientific projects.

These people need vast amounts of reliable hardware that they can trust.

>>791916

This

>>791925

Nice argument.

>>792117

>What makes you think Power isn't backdoored just like x86?

Because we have access to the code of the hardware.

>inb4 but they could have build it in the hardware

Yes tons of papers were already made on the subject.

This is the best that we have for now until we can produced 12nµ CPUs at home.

So fuck of with these CIA level questions.

There's only benefits to go onto this platform.

You have the source code

You participate in the financing of a small company who is OK into letting you OWN what you buy.

What do you want more ?

>>791914

Google is already using Power9 systems, and the newest supercomputers at a couple of the national labs in the US are Power9 systems too.

>>792128

>There's only benefits to go onto this platform.

But it'll drive up my power bill.

>>792136

>But it'll drive up my power bill.

You haven't read the specs.

>>792117

we don't assume that. we simply don't know of any backdoors, unlike x86, which has confirmed backdoors.

>>792171

What if that's what they want you to think? What if they put obvious backdoors into x86 so that you'd think other architectures are safe, but in reality they have the super secret *real* backdoors?

>>792172

>What if they put obvious backdoors into x86

>obvious backdoors into x86

>obvious backdoors

>obvious

https://github.com/xoreaxeaxeax/sandsifter/raw/master/references/domas_breaking_the_x86_isa_wp.pdf

There is so much shit hidden in x86.

POWER9 is part of OPENPOWER.

https://en.wikipedia.org/wiki/POWER9

>the POWER9 architecture is open for licensing and modification by the OpenPOWER Foundation members.

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

>IBM is using the word open to describe this project in three ways.

>They are licensing the microprocessor technology openly to its partners. They are sharing the blueprints to their hardware and software to their partners, so they can hire IBM or other companies to manufacture processors or other related chips.

>They will collaborate openly in an open-collaboration business model where participants share technologies and innovations with each other.

>Advantages via open source software such as the Linux operating system.

AKA this means that people like at raptor engineering are partners and since the license is "Openly shared" To their partners raptor can share the software of the hardware.

Otherwise this project wouldn't have the RYF certification.

>>792172

What if you are a covert NSA propagandist paid by the Rothschild to prevent us from buying non-backdoored computers? What if Satan seeded fossils under the ground so that we would think evolution is real? What if the world was created just 15 seconds ago with all of our memories of the past built-in?

There are countless unfalsifiable scenarios I could come up with, but because we can't know whether they are true (by definition) we don't even bother considering them.

>What if that's what they want you to think? What if they put obvious backdoors into x86

if that was their plan they made such a great job, didn't they. the Intel and AMD backdoors look sooo real that I'm, indeed, compelled to believe they are real.

> ...so that you'd think other architectures are safe

again; I don't think they are safe. All I think is that they are MUCH more likely to be safe, given the state of the evidence.

>>792174

>Otherwise this project wouldn't have the RYF certification.

the RFY certification doesn't require that hardware be built from libre (or even "shareable") designs. it's still too early in the era of free hardware to make such demands.

to be honest I don't know whether Power9 core designs are freely shareable in some sense. What I know is that Talos will be able to function with 100% libre software, and that's fucking awesome enough; because all hardware in that category so far has been either old or slow (compared to high-end desktops/workstations/servers)

Bump again because this is TIME LIMITED and if it fails it would be most unfortunate.

If you got the cash atleast buy a motherboard!

>>790147

>>790724

Aren't these just off the shelf Mobos with just a custom firmware? These CPUs were the latest they would cost like a 2000 dollar Xeon because Power9 is way more niche than Xeon.

>>792889

if these*

>>792889

>Aren't these just off the shelf Mobos with just a custom firmware?

No, they have to develop the actual hardware too, since the only POWER-capable motherboards available right now are not ATX-compatible and whatnot.

>>792889

Where the fuck do you shop that has ATX POWER9 mobos on the fucking shelf?

How does Talos compare to QUBES?

>>792919

nigger...

>>792905

ssi eeb//SSI CEB cases.

https://www.newegg.com/Product/ProductList.aspx?Submit=ENE&IsNodeId=1&N=100007583%20600474033%20600483948

>>791916

>some sort of UNIX

>implying AIX

>implying HP-UX

>implying IRIX

>implying Digital UNIX

>implying Solaris

It ain't the 90s anymore, nobody uses anything other than Windows, Mac, and Linux on workstations. I mean, name one company that actually uses BSD outside backend.

>>793219

nintendo, sony

>>793219

You can use linux on this.

>>793243

Those are even less BSD than Android is Linux. I don't think 99% of BSD installs even have a DE running.

>>793219

>mac user

>saying nobody uses unix

get the fuck out of this board you fucking retarded idiot

do a bit of thinking next time you post anywhere pls

or just hang yourself

https://www.opengroup.org/openbrand/register/

>>793219

OSX is BSD based, and now that Oracle is killing off SPARC and Solaris, the only major Unix 03 compliant OS.

TODAY IS THE LAST DAY!

Good luck to TALOS II

Kinda' off topic, but does anybody have the video of the Raptor guy giving the "game over for general computing" soliloquy?

>>788797 (OP)

>Anyone who is interested in the Talos II?

The more I read up on it, the more I want it to succeed and become a viable player on the market. I can't afford that specific system, hopefully they will succeed and have a budget option later on for people like me.

>>793816

This

> have a budget option later on for people like me.

HAs long the quality doesn't drop down I don't say no for cheaper product.

But if it can't be cheaper like me you should just put some money aside for this kind of gem.

>>793836

It seems cheap now but it's a bit pricey compared to an x86_64 HP Z- Workstation.

Still, I think POWER9 is worth it.

>>793816

Yeah, most people with the budget to fill every slot on that board don't give a shit about freedom anyway. If they made a mini-itx version pcengines would go out of business overnight.

>>793848

Is the hardware that utilize x86 architecture even compatible with power9? I agree with you, It would have been nice if they came out with ITX Mobos.

>>793844

fuck yeah it is

>>793966

>Is the hardware that utilize x86 architecture even compatible with power9?

You should ask yourself this question.

>Is PCIe compatible with other PCIe extension cards ?

>>793966

>Is the hardware that utilize x86 architecture even compatible with power9?

Also beside PCIe compatibility the advantage of these kinds of boards is that other hardware manufacturers can fully take advantage of the non-secret instructions and functionalities of the ThalosII motherboards without having to passes some secret trades thus it limits unfair competition from the CPU/motherboard manufacturer.

>how can unfair competition happen ?

Intel hides instruction in it's cpu, besides CIA spying there's also the hardware market that can ask for better optimization for X specific product.

For example Nvidia could ask a better routing of it's instruction or special instructions in intel CPUs for their brand to have better performance or they need to contact the raptor team to make these improvement on the board and these improvements will be made public since the board is RYF certified aka under a free license.

Once the Talos becomes something a bit more spread and has a market share, the second thing that we need are free/libre GPUs to make the insanity of GPU hype stop.

These possibilities re-opens (yes it opens again since they were closed in the mid-80s) for less useless/fake hardware releases, for great optimization/hack, it might bring back the golden age of computing were nonsense/hype was nonexistent.

>>794099

I hope me will get something better than GPUs. GPU is an outdated marketing term that Nvidia invented, and nothing that makes a modern GPU a GPU would be needed on a platform that does not have to run a ton of legacy Windows games. On a new platform we would be much better of with just a big bunch of FPUs on a card. Which is pretty much what modern GPUs are, except with a bunch of proprietary bullshit in front of them to, supposedly, make them less scary for incompetent game devs.

A very funny email thread was produced on the Dyne Mailing List (Devuan's Mailing List) in response to this. It quite literally began talking about Talos II, then it switched to conspiracy surrounding Intel and AMD (not talking about ME but literally screaching "What IF!!!1!" statements), then it devolved into a pissing contest about technological purism, then there were autistics saying "The human condition is why people can't get along." Yeah, man. Real deep.

You gotta love FLOSS Mailing Lists.

>>794108

>a GPU is just a bunch of VPU

No. It's like saying manycore uarchs are just "a bunch of CPU".

Can it run Debian?

>>794126

Does debian have a powerpc installer?

>>794112

>You gotta love FLOSS Mailing Lists.

More like:

>You gotta love intel/amd/arm shills/detractors

>>794129

You forgot the purism shill.

>>794128

oh, I didn't know it was the ppc64el iso, thanks

>>794124

No, it's like saying they are a bunch of CPU cores. Which is, mostly, true.

>>794185

Arm is only used on embedded devices that I know of.

x86 support is fading. Did you mean x86_64?

>MIPS

Oh hai Tanenbaum and MINIX!

Is there a difference between POWER8-9 and PowerPC?

What is SuperH?

>>794220

>Did you mean x86_64?

That's just an extension of x86, same crap.

>>794220

ARM is actually used in servers somewhat. I know there's at least one or two places making server cpus

>>794126

>>794128

>>794132

So wait.. You're saying any distro that supports powerpc will support this too?

>>794447

You will probably need a ppc64le distro. Most of the standard big-endian ppc/ppc64 stuff is built exclusively for Macintosh systems.

>>794220

>Is there a difference between POWER8-9 and PowerPC?

PowerPC is a cut-down, low-power version of POWER mostly used by Apple. These days, PowerPC is mostly only used in embedded applications. The PowerPC 970 marketed by Apple as the G5, for example, was basically a low-voltage POWER4 with the addition of Altivec and the removal of the second core. They shrunk the die to 90nm in the 970FX, and added the second core again in the PPC 970MP. The older PPC 6xx models had a similar relationship with POWER2 and POWER3.

>>794185

Out of curiosity, why SuperH and not SPARC? Was SuperH ever used in anything besides arcade hardware?

>>794447

I don't think PPC64LE will run on anything but POWER 8 and 9.

>>794547

SH2 is patent-free due to age and there are new clones being built as a result.

>>794555

New clones being built where and for what?

>>794185

Shouldnt RISCV be at the bottom?

>>794559

http://j-core.org/

For cheap, auditable embedded chips like USB controllers or NICs

>>794547

>why SuperH and not SPARC

Because the point of the meme is to put something absurd at the bottom.

>>794554

>Which is probably how it ended up on that list.

I was actually thinking of the Saturn when I added SuperH.

>>788797 (OP)

Just ordered one.

>>795546

They're still accepting preorders?

>>788938

>serial console

>ham radio

>embedded devices

>modems

>printers

>plotters

>industrial devices

holy fuck we have typical reddit style cia niggers even here

Still accepting preorders:

https://secure.raptorcs.com/content/base/products.html

>proprietary broadcom BCM57xx ethernet controller

enjoy your botnet

>>802287

It's getting the RYF certification it can't have a non-free component.

>>802287

>>802309

According to the GuixSD documentation, broadcom does have a WiFi chipset with a libre driver

https://www.gnu.org/software/guix/manual/html_node/Hardware-Considerations.html#Hardware-Considerations

Not sure what the status is on the ethernet controller for this, but apparently not every broadcom thing is botnet

>>795546

>>796998

>preorders

I will buy one when it's out, provided it runs full featured on a deblobbed kernel and the board is well manufactured

>>802287

The drivers for their ethernet cards are Free Software from what I've read. It's the wireless that's the problem. This post sounds like FUD, but if it's not I'd like to see the proof.

>>802328

The driver isn't the problem; the controller firmware is.

Director of research at raptor acknowledging the issue in email:

https://mail.coreboot.org/pipermail/coreboot/2017-October/085242.html

>>802333

https://www.mail-archive.com/search?l=coreboot%40coreboot.org&q=subject:%22%5C%5Bcoreboot%5C%5D+Owner%5C-controlled+POWER9+Talos+II+%5C%5Bwas%5C%3A+Disabling+Intel+ME+11+via+undocumented+mode%5C%5D%22&o=newest&f=1

>This is the only device on the board that requires firmware, but it is behind the IOMMU and the FSF has indicated that we can still obtain RYF with this limitation.

>That being said, we'd love to see a libre firmware created for this controller as it would allow not only this system, but future systems, to be blob-free while using a modern non-Intel GbE NIC...

>The NetXtreme devices appear to be quite stable, and their internal operation has already been partially documented, meaning development of a true open firmware port is at least possible.

The controller is guarded against by the MMU, its firmware is loaded every reset (i.e., replaceable), and development of an open source alternative is possible. Botnet potential low, but stallmanu spergers should be aware.

>>802329

>We want users to be able to upgrade and control the software at as many levels as possible. If and when free software becomes available for use on a certain secondary processor, we will expect certified products to adopt it within a reasonable period of time. This can be done in the next model of the product, if there is a new model within a reasonable period of time. If this is not done, we will eventually withdraw the certification.

>>793183

You better be posting images, if you actually did

Will see if I can work out a loan for the board to start. Would love to build this fucking monster myself.

>>790632

check the spec, there's one on the I/O and one as a pin header on the mainboard

>>793183

Are you not getting CPUs with it?

>>810391

Are you high? You shouldn't get a loan for any computer unless you enjoy being raped with interest.

>>792919

Depends if you choose to download extra RAM

>>793219

One of the french openbsd devs works for a company that has 1000's of openbsd desktops.

>>794220

ARM was the basis for Acorn computers in the 90's. You can still buy RISCOS today and run it on a RPi if you want something different than the bloated Linux desktop environment crap.

>>810625

>buy RISCOS

They give it away for free, don't they?

>>810626

Version 5 is free. Not sure what the differences are with 6. I haven't yet bought any ARM board, but I'm going to get one that's well supported by OpenBSD, so it can dual boot. That will be my next "desktop".

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

>>801535

Nope. Just the main board. Going to get the most expensive part out of the way first.

>>810716

That's good. Why spend $150 more on a bundled CPU when you can spend $340 later for the same thing?