New Raspberry is out:

- Broadcom BCM2837 64bit ARMv8 quad core Cortex A53 processor @ 1.2GHz with dual core VideoCore IV GPU supporting OpenGL ES 2.0, hardware-accelerated OpenVG, and 1080p30 H.264 high-profile decode.
- System Memory – 1GB LPDDR2 / Storage – micro SD slot
- Video & Audio Output – HDMI 1.4 and 4-pole stereo audio and composite video port.
- 64 bit support.
- Connectivity – 10/100M Ethernet, WiFi 802.11 b/g/n up to 150Mbps and Bluetooth 4.0 LE (BCM43143 module).
- USB – 4x USB 2.0 host ports, 1x micro USB port for power / Expansion: 40-pin GPIO header.
- MIPI DSI for Raspberry Pi touch screen display / MIPI CSI for Raspberry Pi camera.
- Power Supply – 5V up to 2.4A via micro USB.

Oblviously this hw is fully compatible with Linux, Android and Windows 10 IoT only for 35$.
May be a good crunching machine for Rosy??

---

i'm thinking it may be easier to get r@h on this, but then it won't be all that fast and it is quite expensive :o :p lol

http://www.intel.com/content/www/us/en/compute-stick/intel-compute-stick.html

i'm thinking it may be easier to get r@h on this, but then it won't be all that fast and it is quite expensive :o :p lol

http://www.intel.com/content/www/us/en/compute-stick/intel-compute-stick.html

Would be interesting to know how it compares in terms of throughput vs cost (including running costs).

---

"dcdc" wrote:

i'm thinking it may be easier to get r@h on this, but then it won't be all that fast and it is quite expensive :o :p lol

http://www.intel.com/content/www/us/en/compute-stick/intel-compute-stick.html

Would be interesting to know how it compares in terms of throughput vs cost (including running costs).

beyond costs, i'd think one of the rather big limitations is the fixed ram size e.g. 2Gb and storage size e.g. 32 gb for those 'stick' computing devices

if one is keen on the small form factor, i'd think those mini-pcs may be more useful/versatile in which u'd need to find those that the dram can be swapped with mass market drams and storage with ssd/hdd drives

e.g.
http://www.intel.com/content/www/us/en/nuc/overview.html
http://store.hp.com/us/en/pdp/desktops/hp-pavilion-mini-desktop---300-020-p-k5g62aa-aba--1

apparently amd is coming right into the ring to compete in this category too :D
http://arstechnica.com/gadgets/2014/04/review-gigabytes-amd-brix-gives-intels-mini-pc-a-run-for-its-money/
http://liliputing.com/2015/06/amds-project-quantum-mini-pc-packs-serious-graphics.html
oops, amd packs an intel in there instead!
http://www.pcworld.com/article/2973970/software-games/exclusive-amds-amazingly-tiny-and-powerful-project-quantum-dissected.html

after a little more goggling, it is rather apparent mini pcs aren't literally cheap, the higher end models cost as much as a desktop pc with a decent processor, the only difference probably is its size, being 'small' relatively

i'm thinking it may be easier to get r@h on this, but then it won't be all that fast and it is quite expensive :o :p lol

http://www.intel.com/content/www/us/en/compute-stick/intel-compute-stick.html

But, whit Raspberry you can create clusters!!
https://www.google.com/search?q=raspberry+cluster&biw=1366&bih=635&source=lnms&tbm=isch&sa=X&sqi=2&ved=0ahUKEwjIqsTw25_LAhUJnRoKHbaLD74Q_AUIBigB

---

"VENETO" wrote:

i'm thinking it may be easier to get r@h on this, but then it won't be all that fast and it is quite expensive :o :p lol

http://www.intel.com/content/www/us/en/compute-stick/intel-compute-stick.html

But, whit Raspberry you can create clusters!!
https://www.google.com/search?q=raspberry+cluster&biw=1366&bih=635&source=lnms&tbm=isch&sa=X&sqi=2&ved=0ahUKEwjIqsTw25_LAhUJnRoKHbaLD74Q_AUIBigB

imho 1 of the issues with clusters is that the networking interconnect could be a bottleneck and the sizes would need to be much bigger and power hungry compared to say a haswell i7 or skylake i7 processor
http://www.zdnet.com/article/build-your-own-supercomputer-out-of-raspberry-pi-boards/

one of the more 'unfortunate' things is that we are probably in the last legs of moores law and further miniaturization and shrinking of the transistor while gaining energy efficiency is reaching the state of *impossible*

imho 1 of the issues with clusters is that the networking interconnect could be a bottleneck and the sizes would need to be much bigger and power hungry compared to say a haswell i7 or skylake i7 processor
http://www.phoronix.com/scan.php?page=article&item=phoronix_effimass_cluster&num=1

First benchmarks seem to be very good here, and here . A little cluster with 5/6 Rasp may be convenient..
Or, using Odroid, but it's more expensive.

---

clusters of arm chips would not match the top tier haswell or skylake gflops v energy requirements (gflops per watt)

e.g.
https://www.pugetsystems.com/labs/articles/Haswell-Floating-Point-Performance-493/

but of course that is AVX / AVX2 ")

and the end of moores law means that it is no longer possible to double that performance on the same chip area (for less energy), and it may soon be *impossible* to even do that - packing ever more transistors per mm square

but of course that is AVX / AVX2 ")

:-P

---

and the end of moores law means that it is no longer possible to double that performance on the same chip area (for less energy), and it may soon be *impossible* to even do that - packing ever more transistors per mm square

The future may be a "combination" of different technologies, like this

---

clusters of arm chips would not match the top tier haswell or skylake gflops v energy requirements (gflops per watt)

This. I had a similar thought and did some rough calcs comparing a cluster of Pis and an overclocked 5820K. The Intel chip won by a factor of about 50 IIRC (on gflops for the equivalent cost).....

This. I had a similar thought and did some rough calcs comparing a cluster of Pis and an overclocked 5820K. The Intel chip won by a factor of about 50 IIRC (on gflops for the equivalent cost).....

Yeap. For sure this kind of hw (Pi and others) is better on astronomical projects (with fft calc) than on Rosy..

---

I've just procured one of these little raspberry pi 3's and take delivery of it today. Main reason for purchasing was just to use as a little toy to experiment with various uses. But I think it will be quite cool to have such a little box crunching my favourite project too!

I haven't seen anywhere on this thread however any mention of if Rosetta will even run. Looking at the applications pages it seems that the ARM version of Rosetta runs only on Android, and the most common and efficient OS's for RPi are Linux based, such as Raspbian (Debian Linux) (I know android is Linux based too but pretty sure they are different in some major way).

Basically, will the Linux version of BOINC download the ARM application? Or will I have to install one of the crumby versions of Android developed for RPi to run Rosetta?

Cheers

Matt

---

Or will I have to install one of the crumby versions of Android developed for RPi to run Rosetta?

A version like this seems to be not so crumby

---

the 'ordinary' arm chips are not quite up to speed yet when compared against say 'lower end' intel models such as atom z3735f. a popular device is perhaps the intel compute stick

a rasberry pi zero is 6 to 10 times slower while a quad core arm a9 is as much as 1/3 to 4 times slower than intel compute stick (Z3735F) where it comes to compute

http://www.phoronix.com/scan.php?page=article&item=raspberry-pi-zero&num=4

it would actually appear that a small computer such as an Intel NUC with a recent i3 processor or equivalent could actually be a decent candidate for running r@h jobs say round the clock while consuming low power

the 'ordinary' arm chips are not quite up to speed yet when compared against say 'lower end' intel models such as atom z3735f. a popular device is perhaps the intel compute stick

a rasberry pi zero is 6 to 10 times slower while a quad core arm a9 is as much as 1/3 to 4 times slower than intel compute stick (Z3735F) where it comes to compute...

The original poster was not talking about a Rpi Zero, but instead an Rpi 3, which has - to quote the OP - "Broadcom BCM2837 64bit ARMv8 quad core Cortex A53 processor @ 1.2GHz with dual core VideoCore IV GPU supporting OpenGL ES 2.0, hardware-accelerated OpenVG...". Significantly more 'oompf' than the Pi Zero. R@home ought to be able to harness that quad-core "and" the dual-core GPU to churn out some reasonable work-unit times.

---

Anyone know if Rosetta will run on the Jetson TX1? There's an interesting article on a cluster of them used for number crunching here.

I bought an Intel Compute Stick and fired up Rosetta on it. It cost £58 ex VAT from Ebuyer and has a Z3735F quad-core atom CPU, 2GB RAM and 32GB flash and Win10 (32-bit for some reason).

It's heading for a RAC of 760 which is really impressive. I haven't measured the power draw yet, but I don't expect much will beat it in terms of throughput to price or throughput to power consumption.

---

I've checked the power consumption now- it's averaging around 3.5W when running 4 threads of Rosetta.

---

Has anyone been able to get the RPi3 working with this project? I have 5 devices ready to go but it looks like they're not downloading any tasks.

If someone has some tips to get this working I would be appreciated.

Thanks,
Bogdan