Pine64 Touts Its RISC-V-Based Single-Board Computer and Soldering Irons (pine64.org) 42
PINE64's August update included photos of their first prototype for its upcoming Star64 single-board computer, "the first true RISC-V single-board computer from us...but as I wrote last month it certainly isn't the last RISC-V piece of hardware you'll be seeing from us."
Just as a short recap: Star64 comes with a StarFive JH7110 64bit CPU sporting quad SiFive FU740 cores clocked at 1.5GHz. The SOC is equipped with BXE-2-32 from Imagination Technologies, which is said to be a solid mid-range GPU. Star64 will be available in two configurations — with 4Gb and 8GB of RAM, similarly to the Quartz64. Both hardware versions include USB 3.0 and a PCIe slot as well as two native Gigabit Ethernet NICs.... Along the long leading edges you'll find PCIe on one end and GPIO on the other. At one end of the board you'll find a digital video output, a double-stacked Gigabit Ethernet port and a 12V barrel plug for power. On the opposite side, you'll find 3x USB 2.0, 1x USB 3.0, an audio jack as well as a power button. There are also two U.FL ports for antennas — one for bluetooth and the other for WiFi....
The Star64 also has an MiPi display output complete with a touch panel (TP) input, a 12V power port, a CSI camera port and an eMMC slot. A micro SD card slot can be found at the bottom of the PCB. Similarly to the RockPro64 and Quartz64, the 12V port on the Star64 can be used for powering other hardware directly from the board — a popular example is powering one or multiple SSDs connected to a PCIe SATA adapter. I'll add that, at least in theory, the Star64 would make a great network-attached storage device because of its SoCs low thermals and idle power. I am looking forward to seeing NAS-focused Linux or BSD* OSes available for the board.
Speaking of software, efforts to support the SoC in Linux have already begun. I've been told that both Debian and Fedora are already being ported to the StarFive JH7110, which is great news. We are certain that many other OSes will follow swiftly — especially once we start delivering the Star64 to interested developers. On the subject of availability: the Star64 will be available in a few weeks time, and will initially be available to developers. Given the interest in the Star64's and the SoC powering I hope to see functional distributions available for the board soon after launch.
The announcement also included an update on their PinecilV2 smart mini portable soldering iron (built with a 32-bit RISC-V SoC). "The Pinecil V2 landed earlier this month and sold out almost instantly. The next production run of the ought to be available soon however..."
The Star64 also has an MiPi display output complete with a touch panel (TP) input, a 12V power port, a CSI camera port and an eMMC slot. A micro SD card slot can be found at the bottom of the PCB. Similarly to the RockPro64 and Quartz64, the 12V port on the Star64 can be used for powering other hardware directly from the board — a popular example is powering one or multiple SSDs connected to a PCIe SATA adapter. I'll add that, at least in theory, the Star64 would make a great network-attached storage device because of its SoCs low thermals and idle power. I am looking forward to seeing NAS-focused Linux or BSD* OSes available for the board.
Speaking of software, efforts to support the SoC in Linux have already begun. I've been told that both Debian and Fedora are already being ported to the StarFive JH7110, which is great news. We are certain that many other OSes will follow swiftly — especially once we start delivering the Star64 to interested developers. On the subject of availability: the Star64 will be available in a few weeks time, and will initially be available to developers. Given the interest in the Star64's and the SoC powering I hope to see functional distributions available for the board soon after launch.
The announcement also included an update on their PinecilV2 smart mini portable soldering iron (built with a 32-bit RISC-V SoC). "The Pinecil V2 landed earlier this month and sold out almost instantly. The next production run of the ought to be available soon however..."
Re: (Score:3, Funny)
Sod off.
Re:..and now to mis-pronounce "SOLDER" (Score:4, Funny)
"Sold off"? WTF does that mean?!
Re: (Score:2)
Go ahead and say aluminum.
Re:..and now to mis-pronounce "SOLDER" (Score:4, Funny)
Aluminiumiumiumium
Re: (Score:2)
Go ahead and say aluminum.
From The Good Place [wikipedia.org], The Burrito [fandom.com] (s2:e12): Say Aluminum [youtube.com]
Re: (Score:3)
Oh, you meant "Americans". Perhaps you should develop better than a grade school understanding of the language's features for differentiating between possession and plurality before you wade into trying to criticize its divergent evolution between hemispheres.
Re: ..and now to mis-pronounce "SOLDER" (Score:1)
"colder", "holder", "bolder", "folder".. "sodder"..
If you're going to butcher a language, at least be consistent.
Re: (Score:2)
Get the fuck out of here.
Re: (Score:2)
What, like "bread", and "read"?
Actually that one's ambiguous - if you think about it for a minute.
Re: (Score:2)
Actually that one's ambiguous - if you think about it for a minute.
Yup ;)
I could have made that clearer by saying:
What, like "bread", "read", and "read"?
Re: (Score:2)
If you're going to butcher a language, at least be consistent.
English butchers itself, quite consistently. In fact, It's consistently inconsistent.
Re: (Score:2)
It's a mish-mash. The English themselves can't agree on pronunciations and spellings over any significant temporal or geographic span, and frankly, they're a minority of speakers of the language at this point anyway.
I work with a company in the UK currently, that has employees all over the damn place, and if I've learned anything from the experience, it's that WW2 didn't end the British Empire. 75 different fucking languages in a country the size of Michigan did.
I e
Re: (Score:2)
I can't say I have ever heard someone pronounce solder as sodder in the US, is that how it is said here or in the UK?
Re: (Score:1)
As an American: I used to actually think that solder and "sauter" -- which is how I heard everyone else saying it -- were actually two different things.
If it makes you feel any better, though, I pronounce it correctly (rhyming perfectly with e.g. "folder"), and greatly dislike the way everyone else mispronounces it.
Re: (Score:2)
... It's hard to walk and talk while eating half a salmon.
Imagination Technologies? (Score:3, Interesting)
Does that mean it will be a closed-source, NDA required, possibly Windows only nightmare of a device?
At least that has been the published past of GPU tech from Imagination Technologies. Google it.
And I gotta ask...will all the firmware for this "product" be OPEN SOURCE and freely downloadable so curious types can check it for backdoors & security holes?
And I looked at the webpage in the link. Cute touch posting their update as a Youtube video. No webpage text to directly quote?
Are webpages that expensive to host? Or is this company just trying to look hip (we gotz Discord & Mastodon), act cool (we duz Telegram), and do some virtue signalling (text posts are lame) to the Gen-whatevers out there?
Yep, I know the target audience for this "product".
Re:Imagination Technologies? (Score:4, Interesting)
And I looked at the webpage in the link. Cute touch posting their update as a Youtube video. No webpage text to directly quote?
Are webpages that expensive to host? Or is this company just trying to look hip (we gotz Discord & Mastodon), act cool (we duz Telegram), and do some virtue signalling (text posts are lame) to the Gen-whatevers out there?
Yep, I know the target audience for this "product".
um... you mustve clicked the wrong link, or maybe forgot how to scroll. its a page on pine64.org with a lot of text, pictures, and even a couple embedded youtube videos. i usually just read the text, havent actually watched one of the summary videos.
they also host most of their infrastructure on their own products. i think the current setup is a cluster of their rockpro64 boards.
Re: (Score:2)
I clicked the link in the article and scrolled as far as I could...which was not much. Perhaps they updated the page after I viewed it?
Still, posting the text does not obviate my concerns over the use of Imagination Technologies notoriously closed source products.
Comparisons to ARM, ATOM, ... (Score:2)
Gimme RAM! (Score:4)
I use a Raspberry Pi 4 computer running Ubuntu Linux for weekend surfing&communication. This thing is not fast, but good enough, and it would be much faster and better with 16GB RAM.
Sadly, there is no 16GB option for Raspberries, and there may not exist one for this new Risc V board either :-(
Re: (Score:1)
Often times they have enough cache memory to be comparable to post-celeron Intel on that metric, but the killer is that even if they are multi-scaler and can do out of order execution they dont do it very well. IPC ends up similar to 20 year old athlons, while the clock rate is also similar to 20 year old athlons...
Re:Gimme RAM! (Score:4, Insightful)
My observation is that these novelty cpus perform poorly also, but I want to be specific.
They aren't a "novelty" anything. They aren't meant for a guy who has 300 tabs open in Chrome and has compiles running in the background. They're meant to be the cheapest easily accessible single board computers. Just because someone is using them incorrectly doesn't mean they are bad. If you need 16gb of ram for your daily tasks then buy something with 16gb of ram. This is engineering 101 here, use the best tool for the job.
Re:Gimme RAM! (Score:5, Interesting)
The beauty of Raspbery PI 4 is its huge eco system including different cases, power supply options and whatnot. For a while I messed around with a Banana PI, which was back then more powerful than the Raspberry PI, until I gave up in frustration due to lack of eco system and lack of operating system support or upgrades.
I can deal with the sluggishness and the frequent swapping of my PI 4 for now, but will likely switch to a more powerful system once these become available in the same form factor and with comparable power demand. That new Risc V board could have been such an option, after all it is quite a bit newer, but with 8GB max it doesn't fit the bill. Continue waiting for Raspberry PI 5 ...
Re: (Score:1)
Re: (Score:2)
That’s true of any architecture now. Solidworks loads as fast today as Autocad did on my 386. Back in the late 1970s you could have a dozen users on a PDP-11 with kilowords of ram.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
It all depends on use case. It is nice to have a desktop grade board that has the ability to have 32-64 gigs of RAM for those Chrome tabs, but for a lot of SBC tasks, something with a lot less in the way of CPU and RAM can do the job intended for it. For example, a Raspberry Pi that has a NTP hat and is part of a pool doesn't really need 32 gigs of RAM. Of course, RAM is nice for I/O caching, but a Pi 0W can handle the same task on a lot cheaper SBC. My OctoPrint server still runs quite well on a Rpi 3B
Re: (Score:1)
> My observation is that these novelty cpus perform poorly
Runs Klipper just fine as a fanless platform to augment 3D printing.
Re: (Score:1)
You know, I'd be happy if I could just lay my grubby mitts on any kind of Raspberry Pi for a reasonable price. It seems the only options are to buy at 400% of the manufacturer's suggested price from a scalper (fuck that noise), or get lucky and find one somewhere in Europe and pay 2-4x of the previous going price when you factor in shipping, taxes, duties etc. etc. Slightly less offensive, but still too much trouble.
RISC-V (Score:3)
Right now, the ONLY purpose of RISC-V is to keep ARM honest. I will buy the Star64 only because that alone is an important purpose. However I am not going to be under any illusion that RISC-V is overall superior -- certainly not on price/perf. At least not yet.
Re:RISC-V (Score:4, Insightful)
The purpose is so you can design products that have an embedded processor, with your whole design and toolchain using open source parts.
The target audience is firmware engineers and system designers. Not users.
So you're partly right; there is no purpose at all to you. It is not, and is not intended to be, and never will be, "overall superior" to anything. There are a bunch of design choices that make implementing the hardware easier; that will not result in overall superiority. Ever. It simply results in improved access for people who want to implement their own processor and have existing software toolchains for it; without buying licenses. And for designers who don't want to have to worry about something like ARM getting bought out and future designs becoming expensive, or no longer available.
That's why there are lots mid-sized RISC-V microcontrollers, and not so many application processors so far. Most of the designers who want this type of product are making small embedded devices where most of the cost or effort is in the software.
Re: (Score:2)
So you're partly right; there is no purpose at all to you. It is not, and is not intended to be, and never will be, "overall superior" to anything.
It might be, if enough of the industry gets behind it, which is conceivable because money always moves the needle. Just not having to pay for licensing could literally make RISC-V surpass ARM through this mechanism... eventually. Look at how Linux is now the most performant mainstream/general-purpose kernel, running on the most hardware, doing/enabling the most work. That happened specifically because of the license, which we know because many major contributors (and PR droids) have told us in no uncertain
Re: (Score:2)
One thing that ARM has going against it for general-purpose computing is the lack of any kind of platform standards for device trees and booting. I can't, for example, get a generic ARM64 linux distro that will boot and install on any ARM device like I can with x86_64. Of course for embedded use and running android vendors don't want any kind of openness so this suits them fine. But it's the primary reason we'll never see widespread availability of generic ARM laptops, for example, except from Apple and
Re: (Score:2)
So you're partly right; there is no purpose at all to you. It is not, and is not intended to be, and never will be, "overall superior" to anything.
It might be, if enough of the industry gets behind it
You wave your hands with maybe, but no. Design decisions were made that simplify implementation, but leave out significant (but difficult) optimizations. It doesn't matter how much adoption it gets, it won't have those optimizations. It can't be superior at technical considerations, only at access. And price is theoretically possible, but unlikely. Though that doesn't matter for the microcontroller use cases, because commodity prices for small ICs means that the component cost is basically an irrelevant par
Does its Bluetooth do BLE AoA / AoD? (Score:2)
Re: (Score:2)
The V is pronounced "five" but it is definitely written "V". You'll have a hard time finding the right libraries if you don't get the name right.
But look, if you're not sure the answer, maybe you shouldn't be looking for a chip and a controller board at all, but a driver and software that uses the PCI carrier version? You can still experiment, but you'd be doing it either in userspace or in a kernel module.
That said, the nRF52833 is a $3-6 with multiple dev board options in the 50-100 range, so there isn't
Re: (Score:2)
The V is pronounced "five" but it is definitely written "V". You'll have a hard time finding the right libraries if you don't get the name right.
Sorry. I was still in anesthetic hangover from a minor surgery. When I'm typing while sub-par the mental automation tends to take over and go phonetic. (Notice that I also didn't get the blockquote closed properly. B-b )
But look, if you're not sure the answer, maybe you shouldn't be looking for a chip and a controller board at all, but a driver and software tha