This is perhaps the least stupid implementation of RISC-V use at home I've seen yet, since at least the rest of the platform can bet swapped for something better.
RISC-V Motherboard For Framework Laptop Pricing Starts At $368 In Early Access
Collapse
X
-
Originally posted by Luke_Wolf View PostUnless you're going to get uppity about some fit and finish issues that come as a result of them being designed so that everything is easily replaceable Framework Laptops are in the really good category.
Comment
-
-
Originally posted by Espionage724 View PostI used a Latitude 5591 as my main computer for years (and today) 8th-gen that works pretty well; the default 1080p eDP screen is tiny at 100% but ideal at 125% (I dock 1080p HDMI 100% usually). Built-in keyboard is fine. Overall it's the best thing I got atm and it works!
I've used C2D with 8GB+ RAM tolerably, and imagine the base $300 Framework RISC-V model would be fine to replace it for my use as long as it can run a desktop Linux distro well, and Fedora being mentioned makes it sound pretty good!
Comment
-
-
Originally posted by ehansin View PostPost is a bit of a tangent, but anyone know how the dispaly and "ergonomics" are on this thing? Not going to get one, just curious.
Had to do some stuff at work on a Windows-based 10th Gen Core (so not that old) Dell Latitude 5410 laptop. Wow, really starting to realize what junk so many laptops are. My 2017 Air I'm writing this on is so much better from an "ergonomics" perspective. To be fair, have an 8th Gen Core series Dell Precision 5530 laptop with a 4K display running Arch and so much more fun to be working in.
Looking forward to more quality low-power (ARM possibly, but okay with X86) laptops that I can run Linux in. Windows really starting to become no fun for me, especially on crappy laptops!
There's a vibrant community of Framework folks DIYing expansion cards, what standard hardware can be swapped in, etc. They even have the full CAD of the shells/etc, all (or nearly all) board design specs/etc all posted to their github.
Comment
-
-
Originally posted by muncrief View PostRISC-V is being held back because all the offerings are way overpriced with incredibly low performance.
Yet there are always myriads of nonsensical excuses as to why.
And so long as that continues, sadly, RISC-V is going nowhere.
The milkv mars, the same soc as this, sells for about 50usd for the 4gb model, the raspberry pi 3 is around 35usd for 1gb of ram, The cpu perf is roughly similar (minus vector stuff because again, that was only ratified in 2021), the gpu is decently better in the mars. the rock pi 4 se sells for around 100usd for 4g ram, the lichee pi4a sells around 134usd for 8g ram, so on and so forth. The prices aren't super high out there, the performance isn't super low, tho having to deal with the lack of compiler optimizations hurt a lot, and again, these don't have things like vector acceleration to begin with.
Risc-v is progressing at insanely fast speeds. Literally nothing is holding it back.
Comment
-
-
Originally posted by Quackdoc View Post
this is a horrendous take, the standard "riscv spec" was ratified in 2019, RVV 1.0, RVV was ratified in 2021, despite this we are now getting cores with cortex A55/A57 class performance, the upcomming P670 cores are advertising better perf per area then A78 cores. Risc-v is progressing at an insanely fast speed. I can't even think of a good thing to equate this to.
The milkv mars, the same soc as this, sells for about 50usd for the 4gb model, the raspberry pi 3 is around 35usd for 1gb of ram, The cpu perf is roughly similar (minus vector stuff because again, that was only ratified in 2021), the gpu is decently better in the mars. the rock pi 4 se sells for around 100usd for 4g ram, the lichee pi4a sells around 134usd for 8g ram, so on and so forth. The prices aren't super high out there, the performance isn't super low, tho having to deal with the lack of compiler optimizations hurt a lot, and again, these don't have things like vector acceleration to begin with.
Risc-v is progressing at insanely fast speeds. Literally nothing is holding it back.
Comment
-
-
I've resisted registering a forum account but there are so many haters and for what? Go now and make a riscv64 instance in qemu, it is just a little slower than what might be called usable.
This mainboard product does not compete with any other Framework mainboard and that is exactly what makes it great. Such a product can be brought to market quickly and it is allowed for this innovation to exist in the world.
For anyone wondering how this DeepComputing mainboard exists for the Framework laptop ecosystem so quickly from a third party mainboard manufacturer:
"Keynote: Making RISC-V Real, Fast! - Yuning Liang, CEO, DeepComputing & Nirav Patel, Framework" https://www.youtube.com/watch?v=cVtPt88rQn0
The performance is noticeably better than qemu riscv64 emulation on a typical home desktop system in 2024.
If you have immediate use for a headless RISC-V architecture general purpose computer and want to put it to work today without faffing around with "maybe, might be, someday" development then yes you can get this same StarFive JH7110 CPU as with the topical DeepComputing mainboard on boards in-stock and ready for your purchase in at least a few products:
StarFive VisionFive2 1.3b ($77 usd)
Milk-V Mars ($49 usd)
Pine64 Star64 ($70 usd)
My favorite is not listed, it is the Milk-V Mars in Raspberry Pi CM4 module the Milk-V Mars CM Lite ($39 usd). I have a Pine64 Star64 that was donated to me and several Milk-V Mars CM Lite. The very same bootloader "firmware" is installed on all these different boards having the same JH7110 CPU and all are running official Debian Linux trixie/testing OS.
The next step-change might be the Eswin EIC7700X CPU to be featured on the following (forward-looking) products:
SiFive Performance P550
Pine64 StarPro64
Milk-V Megrez
Meanwhile you will see a lot of interim low cost development boards and CPUs being peddled to "see what noodles will stick to the wall". Some have more or less cores than the StarFive JH7110, more or less memory, worse or better efficiency and power per core, unfixable security vulnerabilities in silicon, and so on; What we know for sure is that support for JH7110 exists today in upstream Linux and as stated, it is a useful measure of improvement over qemu. Cost may be $30 or $300 and having real hardware saves you time compared to qemu. Given that you are needing to develop RISC-V as is true in academia and many real world applications that nobody will ever know about (because there is no required reporting if you use RISC-V in your company internally as part of some other silicon design) - What is your time worth?
Footnotes:
Debian will not have riscv64 listed on the user-facing website until the first stable release of trixie (currently testing at time of writing this comment) and this is due to not often adding new architectures, and how things happen logically. As of 2024 November the riscv64 weekly and daily installer builds are available along with all architectures in the archive:
U-Boot is the conventional "firmware" for all boards having JH7110 CPU and implements EFI for the operating system being loaded. The trend in U-Boot development is to adopt Linux Kernel as the centralized source of devicetree data. Officially there is no support for this mainboard of topic in U-Boot however the support in mainline Linux is queued for the next release cycle; it is reasonable to say that U-Boot upstream will land support soon. There is no way to make this happen any faster than it already is.
Comment
-
-
The big motivation behind RISC-V from the Industry's perspective is really killing ARM for myriad of reasons much of which relate to their licensing. Which is why Qualcomm, Samsung, Imagination, etc are part of the RISC-V org, and the question is really what's their internal roadmap on modifying their micro-architectures to use RISC-V rather than ARM and at that point there will be some pain in the Android ecosystem but once things are squared away it'll result in a rapid shift in phones, smart TVs, etc at that point. I wouldn't foresee it making serious inroads into x86 space though until maybe a long long time after that.
Comment
-
-
Originally posted by Luke_Wolf View Post
The big motivation behind RISC-V from the Industry's perspective is really killing ARM for myriad of reasons much of which relate to their licensing. Which is why Qualcomm, Samsung, Imagination, etc are part of the RISC-V org, and the question is really what's their internal roadmap on modifying their micro-architectures to use RISC-V rather than ARM and at that point there will be some pain in the Android ecosystem but once things are squared away it'll result in a rapid shift in phones, smart TVs, etc at that point. I wouldn't foresee it making serious inroads into x86 space though until maybe a long long time after that.
Comment
-
Comment