RTX2010 – Radiation-hardened stack machine microprocessor
en.wikipedia.org
RTX2010 – Radiation-hardened stack machine microprocessor
1–10 of 61 posts
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#2Tech specs are available on the manufacturer's site:
http://www.intersil.com/en/products/space-and-harsh-environm...
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#3Philae is controlled by 2 RTX2010 CPUs (hot red), further 8 control the Experiments. 13 CPUs (10* RTX2010, 1* ADSP-21020, 2* 80C3x) in total. © https://twitter.com/philae2014/status/427842417920712704
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#4I wonder how hard it is to create a radiation hardened CPU. Is the premise that radiation can flip bits in the on-die cache and registers? would simply making those redundant be a solution? Or could a single flash of radiation invalidate all memory on the device?
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#5I'm sure Mr. Moore feels vindicated that the fruits of his design work have powered humanity's first comet landing ever.
Still, it's telling of the industry's regard for Forth that the manufacturer itself manages to misspell the language as "Fourth" in the chip's official product page.
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#6I wonder how hard it is to create a radiation hardened CPU. Is the premise that radiation can flip bits in the on-die cache and registers? would simply making those redundant be a solution? Or could a single flash of radiation invalidate all memory on the device?
There's all sorts of problems you have to deal with. Memory flipping is one of the most visible, but that radiation also will damage the lattice of the processor, making some transistors harder to flip on, and some much easier to flip on, leading to transient glitches, etc.
Redundancy's definitely one solution, which is one reason why spacecraft tend to have multiple processors and/or processors with redundant logic pipelines. They also tend to use different substrates that are harder for cosmic rays to effect, different, more expensive, casings that offer more shielding, and also larger, older, more vetted processes. For example, this chip is uses a 1 micron process, which is a lot bigger than the process used in current cpus, which means that there's a lot more mass in the transistors to soak the effects of radiation, making it harder to cause damage.
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#7Hah, an old friend. A company I worked with had a Novix in 1986/1987 for prototyping purposes (early days number plate recognition system). Before we got it we had to sign an unreasonable amount of paper. Rumor had it this is the chip powering the Tomahawk cruisemissile.
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#8I wonder how hard it is to create a radiation hardened CPU. Is the premise that radiation can flip bits in the on-die cache and registers? would simply making those redundant be a solution? Or could a single flash of radiation invalidate all memory on the device?
There's all sorts of problems you have to deal with. Memory flipping is one of the most visible, but that radiation also will damage the lattice of the processor, making some transistors harder to flip on, and some much easier to flip on, leading to transient glitches, etc. Redundancy's definitely one solution, which is one reason why spacecraft tend to have multiple processors and/or processors with redundant logic…
Or we can use alternative semiconductors that are more radiation-resistant. One of many reasons why materials discovery is still very important even if initial chips will never match the performance of silicon.
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#9I wonder how hard it is to create a radiation hardened CPU. Is the premise that radiation can flip bits in the on-die cache and registers? would simply making those redundant be a solution? Or could a single flash of radiation invalidate all memory on the device?
Yes, one of the premises is it flips bits, a phenomena called Single Event Upset(SEU) - e.g. the open source LEON FT (http://www.gaisler.com/index.php/products/processors/leon3ft) processor, is designed to be able to correct and detect such errors - in addition to being manufactured with materials that block radiation. In spacecrafts such radiation hardened processors are often used in addition to having redundancy.
One issue with only using redundancy to overcome errors from radiation is you'll have a hard time determining which processor has an error - especially when several processors are affected at the same time.
Re: RTX2010 – Radiation-hardened stack machine microprocessor
#10Another microprocessor able to execute Forth natively is the b16, from one of the original authors of Gforth.