[HN Gopher] Barcelona Supercomputing Center presents Sargantana:...
___________________________________________________________________
Barcelona Supercomputing Center presents Sargantana: new open-
source RISC-V chip
Author : pimterry
Score : 327 points
Date : 2023-12-14 11:55 UTC (11 hours ago)
(HTM) web link (www.bsc.es)
(TXT) w3m dump (www.bsc.es)
| paulluuk wrote:
| > The BSC, Europe's leading developer of open source computing
| technologies
|
| > The fact that the [..] architecture [..] of these new
| processors is open source, and therefore non-proprietary and
| accessible to all, reduces technological dependence on large
| multinational corporations
|
| I hadn't heard of either BSC nor Open Source Computing before.
| I'm curious though, are there a lot of people out there who are
| not tied to large corporations and who have the knowledge and the
| means to produce computer hardware? Are there hobbyists out there
| producing their own custom chips and graphics cards?
| alfonsodev wrote:
| I don't know about hobbyists but there are less known companies
| doing open source hardware for sure, [1] here is an example of
| cool stackable parallel computing project. I participated on
| the campaign and received mine, but not sure how are they doing
| today, it was a while ago.
|
| Edit: Andreas Olofsson the original founder seems to be still
| active in the field [2]
|
| - [1] https://www.kickstarter.com/projects/adapteva/parallella-
| a-s...
|
| - [2] https://x.com/zeroasic?s=21&t=xSlFhUGn5i8d8RkXrsgAIg
| tecleandor wrote:
| The BSC has been featured a bunch of times around here due to
| their Marenostrum Supercomputer. A month ago someone posted a
| virtual visit to their Marenostrum 4 location, that's kinda
| surprising/interesting because is located inside an old chapel:
| https://news.ycombinator.com/item?id=38160675 https://en.
| wikipedia.org/wiki/File:MareNostrum_4_supercomputer_at_Barcelon
| a_Supercomputing_Center_1_br.jpg
|
| Their Marenostrum 5 is number 8 in the TOP 500 supercomputer
| list ( https://www.top500.org/system/180238/ ) and I think it
| recently started working or it's about to do it now (
| https://www.eetimes.com/bsc-about-to-dispatch-marenostrum-5-...
| ) . They had to change its location as it didn't fit in the
| church anymore, though.
| ciberado wrote:
| But they will keep the Marenostrum 4 in the chapel this time,
| instead of replacing the old generation with the new version
| :).
| tecleandor wrote:
| Nice! The real Computing Church!
| jacquesm wrote:
| If they ever get AGI going it will have come full circle.
| You can go there to pray to your very visible god. Prompt
| engineering will be the new praying, you read it on HN
| first...
| bee_rider wrote:
| Training in the Cloud, fine tuning in old churches,
| inference in your home shrine.
| malwrar wrote:
| If anyone likes ambient music, an artist I like produced an
| album from recordings of marenostrum:
| https://m.youtube.com/watch?v=1EGmWY91Vus
|
| I find it oddly relaxing.
| rwmj wrote:
| They co-hosted the RISC-V Summit back in 2018:
| https://riscv.org/proceedings/2018/05/risc-v-workshop-in-bar...
| kinow wrote:
| For anyone who hasn't heard about the BSC, you can check out
| the website or, if you are more inclined to read code:
|
| - https://earth.bsc.es/gitlab/es/autosubmit/ - project I joined
| last year to work on, a workflow manager used in MareNostrum to
| run mainly (but not exclusively) climate experiments -
| https://earth.bsc.es/gitlab/es/ - other projects from my
| department - https://gitlab.bsc.es/explore/projects - general
| projects
|
| There are also lots of interesting projects, like Aina, a
| project in partnership with Generalitat de Catalunya (like the
| council? prefecture?) to foster the Catalan language with
| models and tools using HPC resources: https://projecteaina.cat/
| MoSattler wrote:
| https://archive.ph/TDj5W
| Y_Y wrote:
| For those of you who don't speak Catalan, "sargantana" is a
| common little local lizard (Podarcis hispanicus, "Iberian wall
| lizard"). Of course the chip family (Lagarto) just means "lizard"
| in Castilian.
| iamsaitam wrote:
| (bonus).. and lagarto is the same in Portuguese as well
| germandiago wrote:
| Warning, offtopic but funny: FWIW "lagarta" in spanish slang
| is a girl with a lot of ambition looking from things from men
| taking advantage of them. Not a "worker" but a dangerous
| person. Lol
| MoSattler wrote:
| same in aragonese
| znpy wrote:
| I don't know, I'm not a lizard expert but that lizard looks the
| same as the ones I saw when I was a kid in the south of Italy.
|
| I guess if I ever start a chip fab there I'm gonna call my chip
| stranvicula or something like that.
| Anduia wrote:
| They are very similar. The Iberian ones are smaller, with
| broader heads, and are sometimes more colorful. I'm pretty
| sure that a Catalan would call the Italian ones 'sargantana'.
| vlugorilla wrote:
| Spanish, castilian does not exist
| dragonwriter wrote:
| Castilian absolutely exists, and us more specific than
| "Spanish".
|
| https://www.merriam-webster.com/dictionary/Castilian
| cosmojg wrote:
| From Wikipedia[1]:
|
| > Castilian (castellano), that is, Spanish, is the native
| language of the Castilians. Its origin is traditionally
| ascribed to an area south of the Cordillera Cantabrica,
| including the upper Ebro valley, in northern Spain, around
| the 8th and 9th centuries; however the first written
| standard was developed in the 13th century in the southern
| city of Toledo. It is descended from the Vulgar Latin of
| the Roman Empire, with Arabic influences, and perhaps
| Basque as well. During the Reconquista in the Middle Ages,
| it was brought to the south of Spain where it replaced the
| languages that were spoken in the former Moorish controlled
| zones, such as the local form of related Latin dialects now
| referred to as Mozarabic, and the Arabic that had been
| introduced by the Muslims. In this process Castilian
| absorbed many traits from these languages, some of which
| continue to be used today. Outside of Spain and a few Latin
| American countries, Castilian is now usually referred to as
| Spanish.
|
| [1] https://en.m.wikipedia.org/wiki/Spanish_language
| Y_Y wrote:
| From the page you linked:
|
| > Name of the language
|
| > In Spain and in some other parts of the Spanish-
| speaking world, Spanish is called not only espanol but
| also castellano (Castilian), the language from the
| Kingdom of Castile, contrasting it with other languages
| spoken in Spain such as Galician, Basque, Asturian,
| Catalan, Aragonese and Occitan.
|
| > The Spanish Constitution of 1978 uses the term
| castellano to define the official language of the whole
| of Spain, in contrast to las demas lenguas espanolas
| (lit. "the other Spanish languages").
| anthk wrote:
| It's the same language. I'm a Spaniard, so I know it
| well. Name it the way you'd like, it can be called
| Spanish, Espanol or Castellano everywhere from Mexico to
| Patagonia, and from The Canaries up to the Pyrenees.
| Narishma wrote:
| It's mentioned in the article.
| ansible wrote:
| Here's a pre-print paper I found:
|
| Sargantana: A 1 GHz+ In-Order RISC-V Processor with SIMD Vector
| Extensions in 22nm FD-SOI
|
| https://upcommons.upc.edu/bitstream/handle/2117/384912/sarga...
|
| RV64GC with a subset of the v0.7.1 vector extension. 1.26GHz
| nominal clock on a 22nm process.
| DeathArrow wrote:
| >22nm FD-SOI
|
| That's kind of not good news. I was hoping for 4nm to have some
| alterative to Intel/AMD/Apple.
| imiric wrote:
| It's unrealistic to expect these chips to compete with modern
| manufacturing standards. Still, it's very impressive the
| progress RISC-V has made in the last few years. It's actually
| a viable option for many projects now.
| bibanez wrote:
| Guess what you need to develop 4nm (spoiler, it's a lot of
| money). There are many applications where 22nm is a good
| tradeoff.
| rwmj wrote:
| You don't do your first experimental test chips on 4nm.
| That's where you get to when you have raised hundreds of
| millions after you've gone through several iterations to
| prove to investors you know what you're doing.
| stefs wrote:
| i guess this chip is not for high end gaming machines or
| servers, but rather cars, industrial machine controlling,
| smart fridges, that kind of stuff. during covid production of
| many appliances ground to a halt because of various chip
| shortages. now what if for some reason asian chips became
| unavailable in europe (wars, natural catastrophes, ...)?
| cheap and easy to build is far more important than high end
| performance here.
| camel-cdr wrote:
| So an in-order core that is slightly faster than rocketchip in
| their benchmarks. That doesn't seem all that exciting, except
| for the vector extension, although they only support a small
| subset of it. Thats sounds similar to spatz [0] and given their
| numbers is slightly faster.
|
| [0] https://github.com/pulp-platform/spatz
| ansible wrote:
| The previous DVINO was a 5-stage in-order, this Sargantana
| core is a 7-stage out-of-order write-back with register
| renaming and a non-blocking memory pipeline.
|
| So it is not a full in-order or a full out-of-order design.
| galangalalgol wrote:
| Are the vector extensions the fixed size ones or the
| originally proposed lanecount agnostic ones? That is the
| aspect of riscv i'm most excited about.
| ksec wrote:
| > in 22nm FD-SOI
|
| That basically implies being Fabbed with Global Foundry inside
| EU ( Germany ).
| 0xDEF wrote:
| >RV64GC
|
| >C
|
| What is the purpose of including the RISC-V Compressed 16-bit
| extension set in what is supposed to be a HPC chip? Most
| embedded/IoT RISC-V implementations include that for obvious
| reasons but why here?
| cmrx64 wrote:
| the same reasons motivating C still apply at HPC: higher code
| density means fewer bits wasted representing redundant
| information, better cache utilization, minimization of memory
| fetch bandwidth, etc.
|
| basically, every metric derived from code size is happier
| when you have 20-30% fewer bits representing it.
| brucehoult wrote:
| If you don't have the C extension then you can't run off the
| shelf Linux distros such as Fedora, Debian, Ubuntu, Arch and
| are limited to what you compile yourself e.g. Buildroot /
| Yocto.
|
| However the actual academic paper says it's RV64G, no C.
| cmrx64 wrote:
| The RVV here isn't compatible with mainstream Linux anyway.
| ansible wrote:
| Thanks for the correction Bruce. I was in a rush (never
| post when you are in a rush, or drunk, or angry) and I was
| so used to seeing RV64GC that I didn't notice the absence
| of the 'C'.
| darksaints wrote:
| That sounds like the perfect high end MCU core. Doesn't say
| what the target use case is, but if it's like other RISC-V
| announcements, they're probably talking about general purpose
| CPUs, in which case those specs are pretty disappointing. It's
| a shame that RISC-V has made so little impact in embedded
| electronics.
| DeathArrow wrote:
| Any benchmarks? Does it compares to Intel/AMD at raw power? Does
| it compares with Apple at efficiency?
| sylware wrote:
| It is in an in-order CPU. It is meant for tasks where
| prediction and robustness are important. More like hard-ish
| realtime stuff in nasty environment (or... security? ahem...)
|
| RISC-V moving forward. Good.
| tecleandor wrote:
| Interesting. It'd be nice to know if they're going to focus on
| HPC loads or hobby/consumer too. I should check to see if I still
| know people around the BSC :P
| _fcs wrote:
| From the preprint [1] it looks like it is not meant for
| consumers. This way, Sargantana lays the
| foundations for future RISC-V based core designs able to meet
| industrial-class performance requirements for scientific, real-
| time, and high-performance computing applications.
|
| 1. <https://upcommons.upc.edu/bitstream/handle/2117/384912/sarg
| a...>
| lifeisstillgood wrote:
| I would love to see a clear roadmap from the EU (not been
| successful searching)
|
| My take on this is
|
| 1. this is less about competitiveness at the cutting edge and
| more about security and economic on-shoring
|
| 2. building chips on-shore at the 40-20nm level massively reduces
| risk, increases the likelihood smaller states can build locally
| and solves for most chip needs
|
| 3. chips we need are rarely the cutting edge AI stuff. The vast
| volume of chios will go in as controllers on screens, USB
| connectors and so on. Building plug and play alternatives will
| give local manufacturers choices, and incentives will help.
|
| 4. the big win is security. Does the CEO of sensitive company,
| the head of security services and the general in charge of
| procurement use keyboards, cpus motherboards and monitors made
| from open source chips manufactured in a trusted nation? What is
| the BOM for the challenger tank - how many chips in there that
| are made by whom and ...
|
| the process is long and arduous and the risks are huge.
|
| But we make tanks from steel other materials made in "favoured
| nations" - surely the same applies to silicon?
| Gravityloss wrote:
| There are projects like Helios: Highly Efficient and
| Lightweight Input/output Open Silicon
|
| https://cordis.europa.eu/project/id/190183836
|
| But AFAIK this is just a small part of large amount of multiple
| projects.
| FirmwareBurner wrote:
| A lot of EU semi research goes on at IMEC in Belgium, but EU
| still lacks the actual means of put any of it into production
| on their own soil. EU fabs have given up going beyond 12nm as
| it was deemed too capital intensive.
| ksec wrote:
| >building chips on-shore at the 40-20nm
|
| >the process is long and arduous and the risks are huge.
|
| Plenty of 28nm+ chips Fabs are inside EU. And more are coming
| online. This isn't a _long_ or _arduous_ process.
|
| Edit: Should have been Plenty of 28nm and above. As the
| original quote state.
| FirmwareBurner wrote:
| _> Plenty of sub 28nm chips Fabs are inside EU._
|
| Which are those "plenty" sub-28nm fabs exactly?
|
| AFAIK only Global Foundries Dresden goes down to 22nm and
| 12nm, and I think that's by far the most cutting edge fab
| currently in EU, making the Ryzen IO dies and other such
| things.
|
| But even TSMC's future Dresden fab starting construction next
| year(hopefully) will start making mostly automotive chips for
| NXP, Bosch and Infineon chips at 28nm and 22nm all the way in
| 2027(!), with plans to go to 16nm and 12nm in the further
| future.
|
| Your view on EU cutting edge semi fabrication seems very
| optimistic.
| wiz21c wrote:
| and TSMC is not exactly a european company...
| FirmwareBurner wrote:
| Of course they weren't gonna export their crown jewels
| outside of Taiwan, the same way how the west didn't
| export their crown jewels to Asia when they did the
| technology transfers for semiconductor manufacturing in
| the '70s, making sure to keep their Asian partners at
| least a node behind.
|
| Well well, how the turn-tables.
| toyg wrote:
| Everything gets out in the end. My Italian hometown had a
| "golden age" of silk manufacturing for a while, thanks to
| bugs smuggled out of China. It lasted for a couple of
| decades and then they were again smuggled out to other
| Italian towns. And then of course you have the nuclear
| shenanigans.
|
| If European countries wanted the tech bad enough, they
| would find ways to get it. The problem is not the know-
| how but the massive investments needed to productize it.
| formerly_proven wrote:
| > massive investments
|
| EU is turning back towards Austerity 2.0: Electric
| Saveroo these days.
| FirmwareBurner wrote:
| _> The problem is not the know-how but the massive
| investments needed to productize it._
|
| Are you telling me the EU, the richest block in the
| world, has less money to spend on fabs than TSMC, as if
| the EU is scrapping for change behind the couch cushions.
|
| If only you knew how much money the EU wastes through
| various useless and vanity projects that accomplish
| nothing except getting certain well connected people
| rich, we could have built 3x TSMCs.
|
| But unlike Taiwan, we're lacking in visionary well
| educated tech leaders, and drowning in clueless
| politicians and established gentrified industry players
| who lobby the funds go to their projects instead.
| qwytw wrote:
| > the funds go to their projects instead.
|
| To me it just seems like relying on government funding to
| drive innovation in sectors where private companies have
| incentives to compete is extremely foolish.
| dataking wrote:
| > Are you telling me the EU, the richest block in the
| world, has less money to spend on fabs than TSMC
|
| That could very well turn out to be the case in practice,
| not for lack of money, but inability to provide the
| promised subsidies according to Financial Times:
|
| https://www.ft.com/content/898454ba-8fc2-4b00-a14f-5f9ee1
| 52d...
| FirmwareBurner wrote:
| Having a company an industry dependent on generous
| subsidies from states is a race to the bottom. TSMC will
| just pit you against other countries on the basis of
| "which one of you is gonna give us more of your tax-
| payers' money and we'll build our fab there"
| toyg wrote:
| _> Are you telling me the EU, the richest block in the
| world, has less money to spend on fabs than TSMC_
|
| I didn't say we don't _have_ the money, but that it 's a
| problem to _commit_ the money. It 's basically the norm
| that EU countries unanimously agree that "something
| should be done" on a certain issue, but then disagree on
| how much it should cost and where the money should come
| from. This gets more and more complicated the bigger the
| cost is (and this is an expensive idea) and the farther
| we are from the regular 7-year-budget process (it was
| last agreed in 2020, so jockeying for big items will
| probably resume in 2025-26).
|
| I don't disagree on the overall lack of vision in
| European political classes (hardly a fault of the EU,
| it's common to basically all countries and all levels of
| government), but even a visionary leader would have to
| work hard to get agreement on such a big project.
| JAlexoid wrote:
| Em.... ASML, a Dutch company, produces the tech behind
| these nodes.
|
| It's a question of supply chains - not tech.
| cduzz wrote:
| My understanding is that a 40nm fab is only economically viable
| if it's spent the first several years of its life producing
| high margin chips.
|
| In other words; the life cycle of a 40nm fab is:
| 1997: start building fab 2000: fab goes online and starts
| producing CPUs 2006: fab upgraded 2012: fab
| switches from CPUs to video and memory controller chip sets
| 2018: fab switches to USB controllers and embedded chips
| 2019: fab offline for 2 months because an antiquated but
| critical part is broken and is only brought back online because
| another similarly old fab went offline and sold off their parts
| 2020: fab shut off because of covid 2021: fab found to be
| a write-off because too many things broke while fab was
| offline.
|
| So if you skip straight past the profitable phase, you end up
| spending billions of dollars to make a fab that makes $0.30
| parts, and it'll never be profitable unless those parts are $10
| each, which in turn makes the product they're in unprofitable.
| FirmwareBurner wrote:
| You are correct. Building fabs today only for fabbing much
| older nodes will not be profitable. You have to target 22nm
| and below otherwise you can't afford to jump in the semi fab
| ring.
| hajile wrote:
| TSMC is building a lot of new 28nm production with plans to
| shut down all their older nodes and move everyone over in
| the next few years.
|
| GlobalFoundries (formerly AMD fabs) created a brand-new
| 22nm planar process specifically for older chips as an
| upgrade to other company's 28nm processes.
|
| Profits seem possible if you approach it the right way.
| throwup238 wrote:
| The math works out a lot better when you're upgrading
| pre-EUV fabs or expanding an existing facility. A lot of
| the gear and setup is mostly the same such as wafer
| cleaning, HVAC and isolation, etc and the local
| challenges to setup and labor have been figured out.
| FirmwareBurner wrote:
| We're talking about different things here. I was talking
| about building new fabs for 28nm nodes and you're talking
| about TSMC upgrading existing fabs from older nodes to
| 28nm production.
|
| Of course upgrading an existing older "sunk-cost" fab to
| 28nm production will be profitable, but not building a
| new one from scratch just for that same older node.
| zozbot234 wrote:
| You will always have pure analog electronics and other
| bespoke things that basically don't benefit from anything
| finer than these nodes. Even for digital chips, it makes no
| sense to use leading edge nodes for very simple logic where
| a lot of the area is just contact pads.
| janekm wrote:
| But you can only really make those profitably for a few
| industries (military, medical, seismic come to mind). The
| EU does have the chip fabs for those industries, of
| course...
| phkahler wrote:
| >> But you can only really make those profitably for a
| few industries
|
| I think it's more like they're only profitable if the
| equipment is already paid for. And even then the margins
| may be low.
| jacquesm wrote:
| It's not about what you can do or can't do. It is about
| what you can do _profitably_ and that 's a completely
| different thing.
| voakbasda wrote:
| I have to wonder if the ability to profit depends
| entirely on the established cartel of semiconductor
| manufacturers. They determine the current prices of chips
| in the marketplace.
|
| If entering that marketplace requires competing with
| them, then I am not sure _anyone_ that is not already in
| the market can ever win. The margins are too low and the
| startup costs are too high.
|
| Government intervention seems to be the only possible
| solution, and that option hardly sounds viable when
| considering that cartel's collective lobbying power.
| jacquesm wrote:
| The capital expense on a new fab is crazy. There may be a
| cartel factor but that usually would work to the
| advantage of the manufacturers, so that doesn't seem to
| be the case here.
| mardifoufs wrote:
| There's no real cartel for older nodes. It's not even
| really possible considering how many fabs exist and how
| many players are operating those older fabs.
| cf1241290841 wrote:
| Number of producers of these fabs is still quite limited
| though.
| cduzz wrote:
| I don't think this is a "cartel of semiconductor
| manufacturers" so much as it's been a "shambolic cluster
| of organizations running crappy old fabs into the ground
| producing cheap chips that were subsidized by a prior
| decade's worth of very expensive products."
|
| I can afford to sell gazillions of chips at $0.08 per
| chip if I'm running a fab I didn't pay to build. I'm only
| (barely) paying for the inputs. When Stan, the last guy
| who understands how to run the widget verifier, or
| Elaine, the last lady to understands how to run the
| polishing machine retire, I'll have to close up shop.
|
| Those $0.08 per chip devices have been absurdly
| subsidized in that a replacement infrastructure to make
| them would require that they cost $10 per device, and the
| ecosystem of things built on $0.08 chips isn't viable in
| a $10 per chip world.
|
| In order to have a fab make $0.03 per unit devices, you
| first have to have the fab spend 10 years making $300 per
| unit devices, regardless of the underlying node size of
| those $300 per unit devices.
|
| Likely you couldn't even go back and make a fab that
| makes large volumes of 60nm-90nm node sizes at all, for
| any amount of money, because the equipment to do this
| (new) hasn't been made in 2 decades and no company is
| willing to invest the money to make new crappy old
| equipment.
|
| It's not a nefarious oligopoly as much as a synchronized
| "run the asset to failure" lifecycle of the
| infrastructure.
|
| How much does it cost to make a 300 year old tree?
| photonbeam wrote:
| > How much does it cost to make a 300 year old tree?
|
| Aside from your main point, I found this an interesting
| thought exercise thinking about cost of air, sunlight,
| soil, water and then 300 years of security
| cf1241290841 wrote:
| >Likely you couldn't even go back and make a fab that
| makes large volumes of 60nm-90nm node sizes at all, for
| any amount of money, because the equipment to do this
| (new) hasn't been made in 2 decades and no company is
| willing to invest the money to make new crappy old
| equipment.
|
| I believe your argument assumes that there is a fixed
| cost to produce even 180nm or 350nm ICs that hasnt
| changed since the first one was produced.
|
| We still need 300 years for a 300 year old tree, but 25
| year old technology might now be relatively easy to build
| if we start from scratch.
|
| What was high tech then might be relatively easy to solve
| now. One example might be https://github.com/circuitvalle
| y/USB_C_Industrial_Camera_FPG... being open source
| instead of a multi year, multi million dollar project.
| cduzz wrote:
| Yes, my argument is that producing at industrial scales
| even chunky nodes requires enormous capital expenditures
| and may be impossible without rebuilding large chunks of
| an antiquated and abandoned supply chain.
|
| Even if it is 10% the cost of making the each of the
| individual components involved in making a relatively
| simple 90nm chip, you're still looking at vast costs.
|
| If you're talking about making 30 chips in a university
| fab, sure, I'll concede that it is "possible" but if
| you're talking about propping up an industry built on
| products that require a herd of standardized "$0.30"
| parts made on legacy 90nm fabs, that ship has sailed.
|
| Update your BOM and recertify or raise your costs by an
| order of magnitude.
| cf1241290841 wrote:
| First off, you are definitively making a very solid
| point, cost for getting mass production right are a
| killer once the institutional knowledge is gone. For
| example, its very visible in the field of battery
| technologies if i am not mistaken. Going from lead to
| lithium was a gigantic task and the inertia going
| forwards hasnt reduced enough at this point.
|
| But realistically this is a matter of going back far
| enough, to lower the cost far enough? 10% are a good
| start but to stick to the topic, physical gyroscopes from
| decades ago are now replaced with MEMS ICs where the
| reduction in cost is magnitudes more then down to 10%. At
| a certain point the reduced cost makes it viable. The
| question is just has it been long enough?
|
| While we wont get 90nm cheap enough, the question is what
| can we do on a hobby level (vs academia)? Because going
| from there (neglectable cost and technological
| requirements) to mass production will at some point be
| cheaper then the cost of setting up reproducible tooling
| for older high tech systems.
|
| I am likely still off with 180nm, but there should be a
| level at which this makes economical sense. A level that
| gets cheaper to reach with technological progress / time.
| cf1241290841 wrote:
| Relevant to mention MEMS (micro-electromechanical
| systems) in this context, which use much older nm tech.
| Be it digital micro mirror devices1 or gyros2. Or
| photo/laser diodes.
|
| Given the physical limitations, as well as the problems
| we have with code base security it might be time to aim
| for cheaper production of something in the region of
| 180nm instead.
|
| Looking at how old much of the standard weaponry used
| today is (TOW 50 years with an actual physical gyroscope,
| Javelin still 25 years3), the demand from the military
| alone should cover the initial cost. Especially if you
| look at the ludicrous prices western countries payed for
| even dumb artillery shells.
|
| 1 Texas Instruments DMD from a DLP projector from
| @AppliedScience https://youtu.be/9nb8mM3uEIc?t=428
|
| 2 Explanation of MPU-6050 from @BreakingTaps
| https://youtu.be/9X4frIQo7x0?t=664
|
| 3 Teardown of both from @lelabodemichel5162
| https://www.youtube.com/watch?v=s7-6hgX7-zQ
|
| Sorry for late edits
| cduzz wrote:
| "But I've got a product that's certified with this part
| that's running on a 40nm process that has these
| specifications that are deeply tied to features of that
| 40nm process; things like voltage ranges and temperature
| tolerances! If you force me to switch to a comparable but
| not identical part at 22nm I'll have to re-certify my
| widget with 18 different regulatory agencies!"
| Someone wrote:
| If those are your needs, you order all the parts you need
| over your product's lifetime up-front or get (= pay for)
| a contract with the manufacturer that makes them promise
| to sell you the parts for X years (they probably wouldn't
| keep producing old parts, but would stockpile enough of
| them to be able to deliver working ones years later)
|
| (Or you prepare for having to go to eBay for working
| parts. https://www.nytimes.com/2002/05/12/us/for-parts-
| nasa-boldly-...)
| Workaccount2 wrote:
| There might be an argument then that it would be worth it
| for the state to take the hit. If shit hits the fan and you
| have zero semi-manufacturing, then you are going to be
| pretty screwed.
| qwytw wrote:
| > If shit hits the fan and you have zero semi-
| manufacturing, then you are going to be pretty screwed.
|
| I don't really understand this claim at all. Chips are
| not exactly fungible, unless you force your local
| companies to use you "state sponsored chips" in their
| products just being able to produce "chips" wouldn't be
| that useful. What are you going to do with them?
| 15155 wrote:
| Guide munitions if needed.
| black_puppydog wrote:
| So the cost of building a fab hasn't come down in the last
| decades, huh? Genuinely asking, is there some^W^W^W what is
| the "uncompressible" cost in fab-fabbing? I'd totally guess
| that staff and the building itself are not it?
| RobotToaster wrote:
| How does an entire semiconductor factory become FUBAR from
| being offline for a year?
| cduzz wrote:
| The example is hypothetical, but complex machines can be
| complex to keep running, and often suffer catastrophically
| when shut down.
|
| If the fab was barely profitable before shutting down, it
| doesn't take much to total it. Fabs are full of machines
| that cost tens of millions of dollars when they were new
| and there are simply no spare parts of vendor support for
| them now, and you can't just swap in a modern replacement.
| Fabs are full of extremely sensitive environments (no dust
| here, acid that will kill you if you touch it there,
| constant temperatures, no humidity, etc). If any of that is
| compromised, it's now just a toxic waste dump.
|
| Again, I have no specific knowledge in this domain, but I
| imagine most of the time the owner's happy enough just to
| walk away from the headache.
| tyingq wrote:
| There's also the brain drain aspect. All the process
| engineers and techs that understood all the various
| "recipes", quirks, etc, of the various machines moved on
| to other work.
|
| A new crew will eventually work it out, but there's a lot
| of trial and error getting to the right bake time/temps,
| spin rpm, etc, etc. Yield and rework suffers while they
| do that.
| TheCondor wrote:
| Not an expert, but there are additional start up costs that
| need to be spent to "start it up." With any significant
| downtime, those could eat up any possible profit unless
| it's a newest technology fab.
| JAlexoid wrote:
| Dust is the simplest example.
|
| Once you shut off the dust extraction, you may just end up
| with too much dust collected in the equipment to make it
| utterly useless.
| BiteCode_dev wrote:
| Not all ventures need to be profitable. The EU may decide to
| take a loss on this solely for strategic reasons.
| qwebfdzsh wrote:
| > strategic reasons
|
| Such as? I can't really think of any benefit besides
| providing jobs and funding for contractors (so kicks backs
| etc.)
|
| Then again it's not particularly surprising, the EU is well
| know for wasting massive amounts of money on all sorts of
| nonsense while ignoring things that actually matter.
| mbauman wrote:
| There's both supply-chain and runtime security.
| cduzz wrote:
| Don't forget the MBAs willing to burn it all down to
| juice the Q2 profits.
| KerrAvon wrote:
| Have you looked at a Pentagon budget lately? It's
| entirely welfare for defense contractors.
| mcbits wrote:
| Sounds like there is a need for investment into innovation
| beyond just building the next-generation fab for $2^x
| billion. Bringing the cost of a new less-advanced fab down
| from $2 billion to $100 million, and then building 20 of
| them, could also be profitable (though less exciting). There
| is a national economy that's actually been growing quite well
| for a few decades now by applying that general idea to other
| industries.
| londons_explore wrote:
| > What is the BOM for the challenger tank - how many chips in
| there that are made by whom
|
| In today's world, it would seem more sensible to just stockpile
| enough of all the components for 5-7 years of tank production,
| knowing that if your enemy tries any evil tricks then you have
| half a decade to figure out how to redesign or make the
| components yourself.
|
| Keep a close eye on anything that looks like an antenna and it
| isn't so bad having the enemy backdooring your chips either.
| jes wrote:
| This has been my take as well. There is a lot of disruption
| in a company when a key part, like the FPGA that serves as a
| communications nexus in the product goes EOL and everyone
| scrambles for a year trying to engineer in a replacement.
|
| Buy enough parts for expected product life, make good use of
| the time you didn't waste on scrambling, and when your
| product is EOL sell any left-over parts on the secondhand
| markets.
| mastax wrote:
| I agree that often the less cutting edge chips are important
| but doesn't the EU already have that handled with ST
| Microelectronics, NXP, Infineon? What's lacking is very high
| end CPU, GPU, high end memory, high end FPGA.
| qwytw wrote:
| > about security and economic on-shoring > increases the
| likelihood smaller states can build locally and solves for most
| chip needs
|
| I'm not sure what does that mean? What specific chip needs that
| would that solve and what benefits would this provide? If those
| chips are not competitive nobody would buy them? So what would
| governments do with them? Stockpile them for the future just
| 'in case'?
|
| The problem is that unlike grain or oil chips are not exactly
| fungible if your military production or other vital industries
| lose access to their current suppliers they wouldn't be able to
| use your slow, outdated and overpriced chips anyway (and
| forcing them to do that under normal circumstances would make
| your products less competitive).
|
| > BOM for the challenger tank
|
| How many other components does the Challenger tank contain
| (IIRC it's not really produced anymore anyway) which are not
| manufactured in the UK? In any case stockpiling necessary chips
| etc. just in case the UK won't able be able to acquire anything
| from the US/Germany/etc. seems like a practical approach than
| trying to develop everything inside the country.
| anonymou2 wrote:
| security, yep! they will run Microsoft Windows, Google
| proprietary javascript, and Whatsapp for "secure" communication
| on these chips!!
| incompatible wrote:
| Is there some reason why you wouldn't be able to run a purely
| open source software stack on it, if you wanted? Does
| Microsoft Windows even run on RISC-V?
| rwmj wrote:
| Is this based off CVA6? That's not mentioned.
| m00dy wrote:
| I am one of the fortunate people who could afford to pay a visit
| to BSC.
| capableweb wrote:
| Since when does it cost money? I'm fairly sure it used to be
| free to visit...
| kh_hk wrote:
| Must be poor phrasing and choice of words I guess. I concur
| it's free to visit.
| manuelabeledo wrote:
| It technically does, if you don't live there.
| kinow wrote:
| It still is. You can just book it with reception directly, or
| if you attend a meeting or conference. Whenever I get friends
| in Barcelona I always invite them over too (anyone that works
| there can request a visitor badge and schedule the visit --
| necessary avoid conflicts).
| cmrx64 wrote:
| What about this chip is open source? As far as I can tell,
| nothing. It frustrates me to no end that closed, secret efforts
| inherit the "open source" branding just because the specification
| they implement is participatory and royalty free.
| ThePituLegend wrote:
| In fact, you can get the RTL here: https://github.com/bsc-
| loca/sargantana :D
| cmrx64 wrote:
| !!! perfect, thank you. I'm annoyed now at myself for not
| having found it...
| gchadwick wrote:
| It's a cool project but I do wish these open source processor
| initiatives targetted more realistic design points.
|
| In particular there's often a desire to push out of order design
| into the micro-architecture where the resulting performance just
| doesn't justify it. In this they're achieving a CoreMark/MHz of
| 2.44 (from the paper here:
| https://upcommons.upc.edu/bitstream/handle/2117/384912/sarga...).
| This is very low performance (on a par with the Arm M0+). Now
| CoreMark certainly isn't the be all and end all of Benchmarks. In
| particular it has very little relevance to high performance
| compute or application cores in general. However it's a useful
| performance smoke test. It is easy to perform well e.g getting
| close to 1.0 IPC for a single issue design such as Sargantana,
| CoreMark doesn't really stress the memory system so a major
| source of stalls that you need to hide latency for just isn't
| there. So if you're not hitting that you've definitely got work
| to do on the microarchitecture. They may well have been better
| off trying to build something simpler and putting more design
| time into improving the performance of the basic
| microarchitecture.
|
| The other crucial aspect that's often overlooked is verification.
| This is a major part of producing a new production quality CPU
| design and it doesn't appear to be discussed in the paper at all.
| Maybe once they've released the RTL they'll also release the
| testbench so you can see what they have done.
| gchadwick wrote:
| Though on the CoreMark benchmark they haven't published the IPC
| achieved. You get a large swing in results depending upon the
| compiler used and switches (For RV32 at least I've found GCC
| out-performs LLVM comfortably).
|
| They do have an IPC number for Dhrystone (another tiny
| benchmark that tells you little about real-world performance
| but you should be able to perform well on), that looks to be
| 0.7.
| phkahler wrote:
| Any of these efforts not performing as well as BOOM may be
| suffering from "not invented here". Its already there and
| getting good IPC. Why not start from that.
| cf1241290841 wrote:
| I believe we might be at the point where supply chain security
| (and code base security) might warrant the question why you
| cant implement something on an M0+.
|
| If you really need higher speeds for reaction time, use an ASIC
| or FPGA. We already do this with USB3 or Ethernet controllers.
| pantulis wrote:
| Unrelated note: BSC is a location in the unapologetically crazy
| HBO series "30 coins" season 2, some cool sequences there
| involving a group visit.
| dtjb wrote:
| As far as data centers go, it's beautiful. Like something you'd
| see in a Mission Impossible heist.
|
| https://my.matterport.com/show/?m=oj5FSKsTt7o
| ashvardanian wrote:
| Does anyone know a decent RISC-V developer kit that one can buy
| in the Bay Area today? Or rent somewhere in the cloud? I want to
| start porting our C libraries to RISC-V.
| LeonM wrote:
| There are plenty, SiFive and MilkV sell boards for example. You
| can also just run emulation.
| ashvardanian wrote:
| There are plenty that exist, but i haven't heard of anyone
| using them or any stores selling them.
|
| Emulation isn't enough. I need to benchmark the libraries.
| Emulation will add significant overhead.
| LeonM wrote:
| > I need to benchmark the libraries. Emulation will add
| significant overhead.
|
| Do not expect good performance from RISC-V processors at
| the moment.
|
| Emulation on a modern X86 CPU will outperform any
| commercial available RISC-V processor at the moment.
| camel-cdr wrote:
| This isn't true in my experiance, especially when dealing
| with the vector extension.
|
| But emulation doesn't offer any usefull performance
| insights anyways, except for maybe dynamic instruction
| count.
| brucehoult wrote:
| > Emulation on a modern X86 CPU will outperform any
| commercial available RISC-V processor at the moment
|
| That's not true.
|
| qemu-user is a little faster than the single-issue HiFive
| Unleashed from 2008, but qemu-system is slower.
|
| Against either the dual-issue U74 cores in the JH7110 or
| the small OoO cores in the TH1520 and SG2042 qemu doesn't
| sand a chance on a core for core basis.
|
| It used to be the case that qemu could win on x86 by
| throwing more cores at the problem, but with the 64 core
| SG2042 in the Milk-V Pioneer that possibility has
| disappeared too -- not to mention that the Pioneer is
| $1500 for chip+motherboard (need to add RAM and storage),
| while a 64 core x86 is $5000 just for the chip.
| camel-cdr wrote:
| It depends, mostly on if you need vector support.
|
| Right now, I'd recommend the canmv kendryte k230 which has a
| C908 rvv 1.0 capable core.
|
| If you can wait a bit, mid/end 2024, I'd go for the vision five
| 3 (or whatever is will be called), as it will have RVA22+V
| (iirc) or for the sg2380 which has SiFive P570s and X280s both
| RVA22+V.
|
| If you don't care about vector, then currently anything based
| on jh7110 should be good.
|
| But if you have the time to deal with very slow execution and
| the potential need to report hardware bugs, I'd consider
| benchmarking on rtl simulation of open source cores. (BOOM,
| tenstorrent-bobcat, XiangShian, ...)
| dataking wrote:
| > The Barcelona Supercomputing Center [...] presented on
| Wednesday the new Sargantana chip, the third generation of open
| source processors designed entirely at the BSC.
|
| > Researchers from other universities and research centres such
| as the Centro de Investigacion en Computacion del Instituto
| Politecnico Nacional de Mexico (CIC-IPN) [...] have participated
| in the development of Sargantana.
|
| So this was designed entirely in Spain but it is also joint work
| with a university in Mexico ;-) Nice project though; I've visited
| BSC and they do a lot of cool work there.
| 999900000999 wrote:
| Very cool, I just got a MangoPi and I'm excited to get some stuff
| running on it.
|
| I imagine RISC-V is the future. None wants to pay licensing fees
| to Arm
| mkehrt wrote:
| So, uh, why's it named after a demon?
| cf1241290841 wrote:
| Shout out to affordable subsidized
| https://en.wikipedia.org/wiki/Multi-project_wafer_service
|
| Be it googles OpenMPW Free Silicon Chip Program
| https://developers.google.com/silicon (still active?)
|
| Or the EU subsidized multi project wafer https://europractice-
| ic.com/schedules-prices-2023/
___________________________________________________________________
(page generated 2023-12-14 23:00 UTC)