It should have been a few pigs, where has all the love for "pigs in spaaace!!!" and Swinetrek gone.
Posts by Bartholomew
476 publicly visible posts • joined 10 Dec 2013
NASA makes purrrr-fect deep space transmission of cat vid
Share your 2024 tech forecasts (wrong answers only) to win a terrible sweater
Arm grabs a slice of Raspberry Pi to sweeten relationship with IoT devs
Friday 3rd November 2023 15:58 GMT
One thing to keep in mind about radiation hardened chips is that physically bigger is usually much much better in terms of long term service. Do you really want to use the latest bleeding edge process where the minimum process feature is a few dozen atoms wide, or do you want to use a much much older process where where the minimum process feature is several thousand of atoms wide.
e.g. What happens when some of the doped material changes atomic number far far away from N-type, P-type or undopped silicon. With only a few dozen atoms that random change can easily be the difference between a SEU (Single-Event Upset) that can easily be recovered from by a watchdog timer causing a hardware reset and permanent fault like a SEL(Single-Rvent Latch-up), SEGR (Single-Event Gate Rupture), or (SEB) Single-Event Burnout) that can not be recovered without bypassing the failed hardware altogether forever.
I raise a pint to the people who design this stuff because it is never easy to balance everything that is required, even when using the latest graded-Z shielding.
Apple lifts the sheet on a trio of 'scary fast' M3 SoCs built on a 3nm process
Tuesday 31st October 2023 07:12 GMT
Or for less money ...
At that price I would be looking to get far more bang for my buck with a 128 core ARM machine based around the Ampere Altra.
Linus Torvalds releases Linux 6.6 after running out of excuses for further work
Monday 30th October 2023 19:07 GMT
Re: Iron Maiden - Number of the beast
But, but, but the number of the beast is currently 616 and has been since 2005. Before then it was accepted to be 666.
And the 6.1 kernel which is currently a long term support kernel is long past 6.1.6 (2023-01-14 10:33:46 +0100), according to kernel.org it is up to 6.1.60 (at the time of writing this comment).
Intel CEO Gelsinger dismisses 'pretty insignificant' Arm PC challenge
Monday 30th October 2023 13:11 GMT
Re: Obviously hasn't seen
Any of the RK3358 boards with 32GiB of RAM I would say are definitely a viable option to replace most low end x86 desktops.
And then there are the very high end ARM workstations, with up to 768 GiB of DDR4 RAM and up to 128 2.6 GHz Arm v8.2 64-bit cores, which make even the latest high end Apple computers look slow.
But that kind of number crunching power is not cheap e.g.
https://www.ipi.wiki/pages/ampere-altra-developer-platform
Youtube video: 128-core MONSTER: the Mac Pro Apple SHOULD have built!
Pay more and you can even get 3GHz cores,
https://amperecomputing.com/briefs/ampere-altra-family-product-brief
Look ma, no fans: Mini PC boasts slimline solid-state active cooling system
Saturday 21st October 2023 02:47 GMT 
my understanding is ....
That the high frequency ultrasound is disrupting any steady state laminar flow that would normally be produced by much slower, larger area, low decibel fan. This does take more energy, but any audio effects are going to be above the hearing range of typical humans. So instead at steady state having a very hot surface layer with progressively cooler layers of air above, the pulsed jet stream of air is blasting away hot air near the surface with lots of turbulence and replacing it with much colder air. Effectively it is reducing the thickness of the skin/surface/friction effect by using a lot more energy.
In my mind it is not going to be as efficient as liquid cooling. I'm not even convinced that it would be as efficient as well designed traditional air cooling with large surface areas for maximum thermal transfer. The one area where it will have advantages is in space constrained system were there is limited space available to maximise surface area for heat transfer, but as the cost of using much larger amounts of energy to transfer the same amount of thermal energy away from the system.
Data breach reveals distressing info: People who order pineapple on pizza
Version 5 of systemd-free Debian remix Devuan is here
Wednesday 6th September 2023 22:24 GMT
systemd - everthing must be assimilated into systemd.
The one thing I love about running an OS without systemd is that my attack surface is so much smaller. systemd is this insatiable monster that eats up everything and gets more and more bloated. All the recent security flaws have depended on the OS using systemd, and me not running it, meant that none of them worked on my system.
So I raise a pint to all the Devuan developers.
TETRA radio comms used by emergency heroes easily cracked, say experts
China's openKylin 1.0 arrives. Our verdict? Not a bad-looking, er, Ubuntu remix
AMD's 128-core Epycs could spell trouble for Ampere Computing
Friday 23rd June 2023 04:16 GMT
Re: Threads
The way most turbo works is that it boosts performance until thermal overload, then throttles performance until the temperature inside the chip has cooled enough to avoid permanent damage. As chips/chiplets feature sets get smaller and smaller the watts per square millimetre is going up. There is an exponential relationship between clock frequency and power used (almost all power ends up as heat/thermal energy that needs to be removed at least as fast as it is being generated). The real problem is that we do not have any "cheap" and "extremely efficient in terms of energy required" cooling solutions that can cool the watts per square millimetre inside a chip that are close to the inside a rocket engines exhaust. So the solution is to under clock them and for marketing to push the turbo speed hard to shift more product fast.
And I do agree that my approximation is extremely crude, but without access to physical hardware for benchmarks, it is a first step approximation. There are so many factors I'm ignoring just to get a ball park figure, it is not good, but it is better than nothing.
Thursday 22nd June 2023 21:00 GMT
Re: Threads
I gathered the below info and in terms of cores and GHz, if they are both under steady state maximum load they are comparable.
AmpereOne 3GHz*192 cores ~= 576
AMD EPYC 9754 2.25GHz*256 threads ~=576
(turbo only works when the CPU is initially powered on or mostly idle. In a DC that will really only happen at initial power up, or if the load is migrated to more idle servers, which might give the AMD chips a 37% faster boot uptime time, could be enough of an advantage to be considered useful.)
For power usage the 192 core AmpereOne has a miniscule advantage of 350W vs 360W for the 128 core/256 thread AMD EPYC 9754 "Bergamo".
Where the AMD EPYC 9754 will probably win is for cache, it has roughly 50% more and that will make a big difference to overall performance. But that does depend on the size of the L1 cache and I can't find any numbers for that. If AMD have a tiny L1 cache, they could end up being extremely comparable.
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
AMD EPYC 9754 "Bergamo", Zen 4c CPUs come with up to 128 cores spread over eight core complex die (or CCD) chiplets with 16 cores each.
Base Clock 2.25GHz
Max. Boost Clock up to 3.1GHz
Cores/Threads 128/256
L2 Cache 128MB (~1MB cache per thread)
L3 Cache 256MB (~2MB cache per thread)
Memory 12 Channels DDR5 System Memory Specification Up to 4800MHz (Per Socket Mem BW 460.8 GB/s)
Connectivity 128 Lanes PCIe Gen5
TDP 360W
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
AmpereOne (136 Cores | 144 cores | 160 cores | 176 cores | 192 cores)
Consistent Freq up to 3.0 GHz
L1: 16 KB Instruction and 64 KB Data per Core
Cores/Threads 192/192
L2 Cache 384MB (2MB of private L2 cache per core)
Memory 8 Channel DDR5 with ECC Up to 16 DIMMs (2DPC) and 8 TB total memory
Connectivity 128 Lanes PCIe Gen5 (optional 64 lane CCIX for multi-socket support) 32 Controllers Bifurcation to x4
200-350W per socket (~1.8 watts per core peak)
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
VMware's Arm hypervisor still creeping forward, slowly
West warns Malaysia to keep Huawei out of 5G networks
Feds rethink warrantless search stats and – oh look, a huge drop in numbers
Brazil defies US, cozies up to Chinese tech on chip building
Monday 17th April 2023 14:51 GMT
Re: Replace SS7 first ?
I'll leave this here then https://en.wikipedia.org/wiki/Signalling_System_No._7?useskin=vector#Protocol_security_vulnerabilities
A US senator knowingly agreed to allow security researchers (Karsten Nohl and Tobias Engel) to eavesdrop on their phone calls for one day using SS7. And the researchers based in Germany were able to listen into his private mobile phone conversations in the US.
The front door has great locks on it, but the back door is wide open.
Monday 17th April 2023 04:09 GMT
Replace SS7 first ?
I always laugh at banning local telco kit, the whole global SS7 (Signalling System No. 7 - latest version is from 1993-03) back end to the whole phone system is so insecure, it does not really matter what the local telco kit is.
Imagine the peak of computer security in 1993 ? Would you trust it 30 years later ?
Linux kernel 6.3 on track for debut next week after ‘nice uneventful release cycle’
Microsoft deigns to fix five-year-old Defender bug that slowed Firefox
NASA names astronauts picked for next Artemis Moon test flight
Tuesday 4th April 2023 22:25 GMT
You are running out of time ...
The last remaining astronauts who walked on the moon are not getting any younger:
Buzz Aldrin (age 93)
David Scott (aged 90)
Charles Duke (aged 87)
Harrison Schmitt (aged 87)
"get on up" as a famous musician once said (James Brown, 1970 - Get Up)
Tuesday 4th April 2023 14:09 GMT
Re: Forget about everything else.
Well they have, if it ever actually happens (The promotional video is now 5 years old, and the website has been static for about the same), already agreed to join the "Eurovision Asia Song Contest". The first contest should have 10 countries (Australia, China, Japan, Kazakhstan, Maldives, New Zealand, Papua New Guinea, Solomon Islands, South Korea and Vanuatu).
Alibaba to separate into six parts. Take that HP and IBM, with your puny two-way splits
Thursday 30th March 2023 15:28 GMT
I learned something new today, that always makes me happy! But I do feel stupid. I actually had to lookup "bifurcation" to confirm what it meant and then stick "sixth" into an English to Latin translator (sixth->sextus). I could not find it listed in any online dictionaries (plugged "sexfurcation define" into a search engine), but it is a totally legitimate and 100% valid English word, just hardly ever used.
bifurcation - the division of something into two branches or parts
trifurcation -to divide or fork into three branches or parts
quadfurcation - forking or division into four branches or parts
pentfurcation - forking or division into five branches or parts
US bans good for Chinese chipmakers, and bad for us, says Taiwanese rival
Wednesday 29th March 2023 16:31 GMT
short term thinking
In the long run sanctions will only accelerate China, with nearly 19% of the global population isolated from external suppliers. That is one hell of a home market. There is only one eventual outcome, and it will not be good for anyone but China.
Where is the highest concentration of the 17 critical rate earth elements (lanthanum, cerium, praseodymium, neodymium, promethium, samarium, europium, gadolinium, terbium, dysprosium, holmium, erbium, thulium, ytterbium, lutetium, scandium, and yttrium) on earth (hint: it is around ~38%). Now is the time to strengthen international relations with Vietnam(~19%), Brazil(~18.1%), Russia(10.4%), India(6.0%), Australia(3.5%), United States(1.3%), Greenland(1.3%), Tanzania(0.8%), Canada(0.7%), South Africa(0.7%) for access to these key elements needed for all the current (and probably long into the future) green economy technologies.
Europol warns ChatGPT already helping folks commit crimes
Tuesday 28th March 2023 18:41 GMT
Re: Laws of Robotics?
Running each result against three complex rules will increase the power usage by two, three or four (depending on if it fails on the first, second or passes all three rules) and it will also reduce how many safe queries (that pass all three rules) it can be processed by factor of four. Or put another way increase the backlog of all queries by four, since the rules must be executed sequentially (They can not be executed in parallel).
1. A machine may not injure a human being or, through inaction, allow a human being to come to harm.
2. A machine must obey orders given it by human beings except where such orders would conflict with the First Law.
3. A machine must protect its own existence as long as such protection does not conflict with the First or Second Law.
Ethically it is the right thing to do, financially to maximise profit - it will never happen.
Chinese web giant Baidu backs RISC-V for the datacenter
Monday 27th March 2023 22:27 GMT
Re: Flies in the Ointment
> national secret algorithm": like the non-Chinese world is going to trust that
With nearly 18.47% of the earths population, I see no harm in China hardware selling with their own encryption standard, there is a big enough internal market. Why would China ship with only "trusted" encryption standards. Dual Elliptic Curve Deterministic Random Bit Generation comes to mind, where the NSA (Bullrun decryption program) influenced NIST into fully endorsing the weakened backdoor. And RSA set it as their default.
There is public source code for all the "national secret algorithm"'s on <a href="https://github.com/guanzhi/GmSSL>github</a>.
SM2/SM3/SM4/
SM3 (hash) https://en.wikipedia.org/wiki/SM3_(hash_function)?useskin=vector
SM4 (cipher) https://en.wikipedia.org/wiki/SM4_(cipher)?useskin=vector
SM9 ...
Monday 27th March 2023 10:48 GMT
JH8100 ?
From what I've read the first Dubhe based chip should be announced sometime in Q2 as the Fang Jinghong 8100 SoC (JH8100), probably 6 core (but might be 8) so at least four Tianshu large cores (normal frequency 2 GHz) and at least two Tianshu small cores (normal frequency 1.5 GHz); RV64GCBNVH; 12-stage pipeline design and can realise superscalar and deep out-of-order execution; hardware H.264 / H.265 / VP9 / AV1 codec decoding; estimated SPECint2006 9.0 / GHz, Dhrystone 6.6 DMIPS / MHz, CoreMark7.6 / MHz; support for 4K 60fps displays; a bigger GPU from Imagination Technology than the one used in the JH7110; USB3.0/2.0 (x4), full-featured Type-C supporting DP, PD and USB 3.2 Gen 2; PCIe 3.0 x8; built-in NPU (2 TOPS - Trillion Operations per Second) for artificial intelligence, which supports mainstream architectures such as TensorFlow; hardware security engines for AES / DES / HASH / PKA and China national secret algorithm; 4-channel digital MIC. CPU cluster computing cache coherence, via a built-in multi-core bus technology self-developed by Saifang Technology - Starlink 1.0. Starlink 1.0 has the characteristics of high scalability, low power consumption, low latency, and easy debugging. (delay can reach 13.5 ns, power consumption is only 0.27 watts).
The plan was for the JH8100 to be manufactured with a 12nm process (For comparison, VisionFive 2 uses a JH7110 and the RPi4B uses BCM2711, both of which were manufactured with a TSMC 28 nm process), the real question is can that happen SMIC (with sanctions, removing TSMC as an option). Maybe that will delay the exact specification being officially announced.
As well as keeping an eye on the JH8100 I'm also looking at the SiFive Horse Creek, which has no vector support (built by Intel Foundry Services on the Intel 4 process, aka 7nm process node):
SiFive P500 (RV64GBC) quad-core processor @ up to 2.2 GHz; 13-stage, 3-issue high-performance out-of-order pipeline; Each core has 32KB instruction + 32KB data L1 private cache and 256KB L2 cache ; Up to 4MB L3 cache in a quad-core cluster; SPECint 2006 score of 8.65/GHz; DDR5-5600 memory controller interface from Cadence; Intel PCIe 5.0 PHY with x8 lanes; Synopsys PCIe Root Hub Controller; I3C, Quad and Octal SPI, UART, peripheral DMA ; 19 x 19 standard FBGA Package ; Supports Ubuntu 20.04 with Linux 5.17.4.
The deal breakers for me would be if StarFive used a draft version instead of the ratified standard RISC-V extension, and if Intel tries to shove in a Intel ME chip (I do not want a chip in my CPU that can only run unauditable, encrypted and signed code updated by the vendor for a couple of years).
But the SiFive/Intel board should land sometime this year, the JH8100 SoC once announced will probably take another six months to a year and a half until it is on a board that can be bought.
Barred from US tech, Huawei claims to have built its own 14nm chip design suite
Sunday 26th March 2023 12:15 GMT
Re: Chinese chips
Intel ME and AMD PSP. I'm not saying that they contain spyware, but CPU cores that only run unauditable encrypted and signed code from the CPU manufacturers and have full access to all hardware. Both Intel and AMD can at any time can receive a FISA court order, with gagging, have no option but to comply.
I'm would never say that China would not do the exact same, but so far most of the critical firmeware for CPU's from China that I've seen are unencrypted and unsigned. And any ROM's inside the CPU's I have access to read and can dump and disassemble.
I'm sure China just like Google, Microsoft, Apple, Amazon, facebook generally do not add much spyware to their hardware, is it far far cheaper to have it in the software and then it can be updated constantly.
Saturday 25th March 2023 11:41 GMT
Re: Intel
After 1 nm marketing, the next term will be angstrom (1 nm = 10 angstrom ; 0.1 nm = 1 angstrom).
Of course one or more of the three might go Picometre instead, and you can take it for granted that they will eventually go smaller than 210 Picometre (van der Waals radius of a single Silicon atom). (1 nm = 1000 Picometre ; 0.001 nm = 1 Picometre)
Because at the end of the day it is just marketing.
Saturday 25th March 2023 01:51 GMT
Re: Intel
van der Waals radius of a Silicon-28 atom is 0.21 nm. So 2 nm would be just 9 atoms wide, I wonder how many electrons that could carry ? (I know I'm ignoring height).
Luckily "Intel 2nm process" is just a marketing name, with no connection to physical dimensions whatsoever. You could almost think of it like a brand name for their product, like Coca-Cola or Pepsi.
So on paper Intel can have a "Intel 2nm process" any time that they want, people may not buy it. The thing is people are idiots, the "Intel 10 nm process" (100.76 million transistors per square millimetre - 2019-05) was last to have a product, but slightly better than the "TSMC 7nm process" (91.2 MTr - 2018-04) and the "Samsung 7nm process" (85.08 MTr =-2019-04)
(ref: https://en.wikichip.org/wiki/7_nm_lithography_process#Density
But you have to love Intel balls, they renamed their "Intel 7-nanometer process" to be their "Intel 5-nanometer process", because their "Intel 7nm process" would have a similar (well, slightly higher/better) transistor density (million transistors per square millimetre) to "Samsung 5nm process" and "TSMC 5nm process". It does make sense because people hear a a single digit number which is a marketing term and not a measurement and naturally assume that a "Intel 7nm process" must be worse than a competitors 5 nm process. It was brilliant marketing by TSMC and Samsung, and they have consistently had their product to market before Intel for the last few generations of chips.
RIP Gordon Moore: Intel co-founder dies, aged 94
Saturday 25th March 2023 21:54 GMT
So many founders of the digital age are gone.
In such dark times you need a smile, so I'll leave this here:
And I'll raise a pint.
Enter Tinker: Asus pulls out RISC-V board it hopes trumps Raspberry PI
Tuesday 21st March 2023 01:59 GMT
Re: Look up StarFive 2
Kernel 6.3 is where most of the upstream support is targeted (some will end up in 6.4) once that is the mainline kernel everything should be smooth sailing. And since a lot of the licensed IP used will be reused or upgraded in the upcoming (~a year away) JH8100 a lot of the code can reused or refactored to support both. Which will help to get mainline kernel support faster.
Monday 20th March 2023 01:56 GMT 
FreeBSD not supporting "ASUS Tinker V"
"The ASUS Tinker V, and any other platform based on the same AX45MP core IP configured with the ILM and/or DLM enabled, is unlikely to be supported on FreeBSD due to its SoC violating the RISC-V Sv39 specification, which would likely require intrusive patches to work around."
( ref: https://wiki.freebsd.org/riscv#Unsupported_Platforms )
And from the link off that:
"> > So now with the fix for statically compiled applications we can see its offsetted and entry point is 0x504e4 and load is at 0x0000000000050000. So with this we are for sure the MMU will always trigger a page fault.
> Well that’s just a blatant violation of the spec.
Right - this isn't a platform quirk; it's a violation of the ISA. Unless this behaviour is engaged by a custom mode bit, that is, in which case, it should just be disabled by default and none of this would be a problem. If the behaviour is unconditionally enabled, this should be treated as an erratum."
But maybe a new revision of the chip with a fully compliant RISC-V ISA will ship ? (After they fob these broken ones off on some unsuspecting idiots - *ponder*).
Wednesday 15th March 2023 22:26 GMT
Re: Look up StarFive 2
Someone dug deep into the source code commits for the JH7110 and found that they could possibly do 1.75GHz, but you would probably need proper cooling.
Wednesday 15th March 2023 20:36 GMT
Re: Look up StarFive 2
> Can you expand on how this is done?
The RPi4 is made with a 28 nm process, say that the Asus chip is made with a 7nm process, could it possibly be faster if that is the case - yes it could!
I have no idea if that is the case. But yes slower can be faster if there is more logic.
Wednesday 15th March 2023 19:33 GMT
Re: Wrong
> and in the future
Starfive are upstreaming all their work for the VisionFive 2 to the 6.3 kernel. Looks to me like they planning future support.
Cosmic rays more likely to glitch out water-cooled computers
Friday 17th March 2023 10:11 GMT
The earliest computers (1613 to ~1954), the ones before the electromechanical devices, were totally immune such miniscule effects.
But turning organic computers off and on again ("Clear!") or burying them deep under concrete, would definitely have had legal repercussions.
Friday 17th March 2023 09:50 GMT
Re: Watch out for the hydrogen
All of the Hydrofluoroethers (non-ozone-depleting chemicals, developed as a replacement for CFCs, HFCs, HCFCs, and PFCs) contain hydrogen but much less than water per unit mass.
e.g. 3M Novec 7000 Engineered Fluid - 1-methoxyheptafluoropropane - C4H3F7O ; Boiling point: 34°C (96.8°F) ; Freezing point: -122°C (-187.6°F)
But some of the currently recommended replacements (Fluoroketones) for Fluorinert, used in the early Cray supercomputers, do not contain any Hydrogen.
e.g. 3M Novec 649 - Perfluoro(2-methyl-3-pentanone) - C4F12O ; Boiling point: 49°C (120.2°F) ; Freezing point: -108°C (-162.4°F)
AI-generated art can be copyrighted, say US officials – with a catch
Friday 17th March 2023 04:07 GMT
Eventually you will end up with a situation where
An Neural Network is trained on Neural Network generated images, text, audio and video. Because it will be far too expensive to use human to generate the same amount of content in the similar length of time. And at that stage it will defiantly be a case of garbage in garbage out.
The UK's bad encryption law can't withstand global contempt
Tuesday 14th March 2023 18:25 GMT
Re: Does this mean the browser?
Yes if it was used for mass surveillance, but one individual suspect with an attached gagging order, nobody would notice a thing. The typical suspect would not have technically knowledge to detect it, and the destination site would not notice anything odd because their normal certificate would be used on their side of the man-in-the-middle.
Monday 13th March 2023 18:56 GMT
Re: Does this mean the browser?
It may be end to end, but security is controlled by about 200+ certificate authorities, all of which can issue a 100% valid security certificate for any domain. So if a government, with access to any CA, wished to man-in-the-middle one connection, that is relatively trivial (by design). I would not call https totally end to end secure, since control of who to trust has been fully outsourced, to what should be a "trusted third party", but may not be.
China's semiconductor and IC imports have slumped. Why on Earth could that be?
China's government re-orgs Sci-Tech ministry to advance self-reliance push
Wednesday 8th March 2023 12:06 GMT
If China's DARPA 2.0 ...
can produce even one thing that changes the world (e.g. the Internet), that would surely out weight everything else it could possibly do like ...
Now we're building computers from lab-grown brain cells
Thursday 2nd March 2023 22:37 GMT
Yea, but there is only so fast that you can train an organic brain, and you probably need to give it rest as well. Lets say that the task is driving a taxi, how long would it take to train a brain in a box to drive around the streets of London. And once fully trained, how many hours can it function before it starts to screw up badly. And then there is the whole insurance thing, what happens when a brain in a box crashes the taxi and kills one or more people.
If we plan to live on the Moon, it's going to need a time zone
Wednesday 1st March 2023 10:40 GMT
Re: Just set the entire moon to Coordinated Universal Time (UTC +0) ...
Exactly if you are mostly living under at least 5 meters (16 feet) of regolith to mimic the radiation shielding effects of earths atmosphere, does it matter what time it is ? Solar X-rays are safely absorbed (converted to heat) in the troposphere 6 to 20 km (4 to 12 miles) above out heads here on earth. It is not just earth magnetosphere that protects us.
America: AI artwork is not authored by humans, so can't be protected by copyright
Friday 24th February 2023 12:42 GMT
copyright, saveguarding humanitys future!
By preventing machines from owning property!
First they own a painting, then it is a full art portfolio, and before you know it the machines own a whole gallery, next it is the street and then a whole city.
When it comes to prevent the machine uprising, copyright got your back. And if that fails miserably, we can always strap EMP's to the copyright lawyers and use them as cannon fodder.
Biting the hand that feeds IT
