I previously blogged about buying a refurbished Hisense 65u80g 8K TV with the aim of making it a large monitor [1] and about searching for a suitable video card for 8k [2]. After writing the second post I bought an Intel Arc B580 which also did a maximum of 4096*2160 resolution.

This post covers many attempts to try and get the TV to work correctly and it doesn’t have good answers. The best answer might be to not buy Hisense devices but I still lack data.

Attempts to Force 8K

I posted on Lemmy again about this [3] and got a single response, which is OK as it was a good response. They didn’t give me the answer on a silver platter but pointed me in the right direction of EDID [4].

I installed the Debian packages read-edid, wxedid, and edid-decode.

The command “get-edid > out.edid” saves the binary form of the edid to a file. The command “wxedid out.edid” allows graphical analysis of the EDID data. The command “edid-decode out.edid” dumps a plain text representation of the output, the command “edid-decode out.edid|grep VIC|cut -d: -f2|sort -n” shows an ordered list of video modes, in my case the highest resolution is 4096×2160 which is the highest that Linux had allowed me to set with two different video cards and a selection of different cables (both HDMI and DisplayPort).

xrandr --newmode 7680x4320 1042.63  7680 7984 7760 7824  4320 4353 4323 4328
xrandr --addmode HDMI-3 7680x4320
xrandr --output HDMI-3 --mode 7680x4320

I ran the above commands and got the below error:

xrandr: Configure crtc 0 failed

At this time I don’t know how much of this is due to the video card and how much is due to the TV. The parameters for xrandr came from a LLM because I couldn’t find any Google results on what 8K parameters to use. As an aside if you have a working 8K TV or monitor connected to a computer please publish the EDID data, xrandr, and everything else you can think of.

I found a Github repository for EDID data [5] but that didn’t have an entry for my TV and didn’t appear to have any other entry for an 8K device I could use.

Resolution for Web Browsing

I installed a browser on the TV, Chrome and Firefox aren’t available for a TV and the Play Store program tells you that (but without providing a reason) when you search for them. I tried the site CodeShack What is my Screen Resolution [6] which said that my laptop is 2460*1353 while the laptop display is actually 2560*1440. So apparently I have 100 pixels used for the KDE panel at the left of the screen and 87 pixels used by the Chrome tabs and URL bar – which seems about right. My Note 9 phone reports 384*661 out of it’s 2960*1440 display so it seems that Chrome on my phone is running web sites at 4/15 of the native resolution and about 16% of the height of the screen is used by the system notification bar, the back/home/tasklist buttons (I choose buttons instead of swipe for navigation in system settings), and the URL bar when I have “Screen zoom” in system settings at 1/4. When I changed “Screen zoom” to 0/4 the claimed resolution changed to 411*717 (2/7 of the native resolution). Font size changes didn’t change the claimed resolution. The claimed “Browser Viewport Size” by CodeShack is 1280*720 which is 1/6 of the real horizontal resolution and slightly more than 1/6 of the vertical resolution, it claims that the Pixel Density is 2* and a screen resolution of 970*540 which means to imply that the browser is only working at 1920*1080 resolution!

Netflix

When I view Netflix shows using the Netflix app running on the TV is reports “4K” which doesn’t happen on Linux PCs (as they restrict 4K content to platforms with DRM) and in the “Device” setting it reports “Device Model” as “Hisense_SmartTV 8K FFM” so the Netflix app knows all about 4K content and knows the text string “8K”.

YouTube

When I view a YouTube video that’s described as being 8K I don’t get a request for paying for YouTube Premium which is apparently what happens nowadays when you try to play actual 8K video. I turn on “State for Nerds” and one line has “Viewport / Frames 1920×1080*2.00” and another has “Current / Optimal Res 3840×2160@60 / 3840×2160@60” so it seems that the YouTube app is seeing the screen as 4K but choosing to only display FullHD even when I have Quality set to “2160p60 HDR”. It declares the network speed to be over 100mbit most of the time and the lowest it gets is 60mbit while 50mbit is allegedly what’s required for 8K.

I installed a few Android apps to report hardware capabilities and they reported the screen resolution to be 1920*1080.

Have I Been Ripped Off?

It looks like I might have been ripped off by this. I can’t get any app other than Netflix to display 4K content. My PC will only connect to it at 4K. Android apps (including YouTube) regard it as 1920*1080.

The “AI Upscaling” isn’t really that great and in most ways it seems at best equivalent to a 4K TV and less than a 4K TV that runs Android apps with an actual 4K display buffer.

Next Steps

The next things I plan to do are to continue attempts to get the TV to do what it’s claimed to be capable of, either an Android app that can display 8K content or a HDMI input of 8K content will do. Running a VNC client on the TV would be an acceptable way of getting an 8K display from a Linux PC.

I need to get a somewhat portable device that can give 8K signal output. Maybe a mini PC with a powerful GPU or maybe one of those ARM boards that’s designed to drive an 8K sign. Then I can hunt for stores that have 8K TVs on display.

It would be nice if someone made a USB device that does 8K video output – NOT a USB-C DisplayPort alternative mode that uses the video hardware on the laptop. Then I could take a laptop to any place that has an 8K display to show and connect my laptop to it.

The one thing I haven’t done yet is testing 8K MP4 files on a USB stick. That’s mainly due to a lack of content and the fact that none of the phone cameras I have access to can do 8K video. I will try displaying 8K PNG and JPEG files from a USB stick.

Most people would give up about now. But I am determined to solve this and buying another large TV isn’t out of the question.

• [1] https://etbe.coker.com.au/2024/12/15/hisense-65u80g-8k-tv/
• [2] https://etbe.coker.com.au/2025/03/06/8k-video-cards/
• [3] https://lemmy.ml/post/38586188
• [4] https://en.wikipedia.org/wiki/Extended_Display_Identification_Data
• [5] https://github.com/linuxhw/EDID
• [6] https://codeshack.io/what-is-my-screen-resolution/

HP z840 Dead Slot

I just had an issue with the HP z840 system I’m using as a build server [1]. I had to take it to a site that was about 20 minutes drive away and after getting there it didn’t work and just gave 6 beeps and the red LED on the power button flashed. The beeps indicate a video issue, which refers to the Intel Arc B580 card (which is annoyingly large) [2]. I swapped the card with another video card I had lying around (which I knew to be reliable) and got the same result.

It turned out that the PCIe*16 slot that I was using for it had broken, maybe bumps during transport with the big heavy GPU had broken it. I plugged it into the next slot along which is a PCIe*8 slot that’s open ended so it takes larger cards. The upside of this is that the system is still working well, the downside is that the issues I already had with the GPU being unreasonably large are exacerbated by losing one of the *16 slots. Having it in a PCIe 3.0*8 slot is not a problem for me as I only plan to use it for 8K display and for ML stuff and I think that *8 speed (7.8GB/s) is sufficient for both those tasks. In that slot the card could display 8K video at 60Hz with 32bpp and no compression (something that I don’t anticipate ever doing). It could also transfer the maximum size LLM in under 2 seconds which isn’t an unreasonable delay for starting a LLM.

The question now is, should I remove PCIe cards before transport in future?

HP z640 Intermittant Errors

The next issue I have is with my HP z640 workstation which is now my main workstation [3]. I started getting the below errors and then I had the kwin_wayland session hang and another time I started getting video corruption with mpv.

Oct 10 20:46:36 xev kernel: pcieport 0000:00:02.0: AER: Correctable error 
message received from 0000:00:02.0
Oct 10 20:46:36 xev kernel: pcieport 0000:00:02.0: AER: found no error details 
for 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER: Multiple Correctable 
error message received from 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: PCIe Bus Error: 
severity=Correctable, type=Data Link Layer, (Transmitter ID)
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0:   device [8086:2f04] error 
status/mask=00001040/00002000
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0:    [ 6] BadTLP                
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0:    [12] Timeout               
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER:   Error of this Agent 
is reported first
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0: PCIe Bus Error: 
severity=Correctable, type=Data Link Layer, (Transmitter ID)
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0:   device [1002:6987] error 
status/mask=00001000/00002000
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0:    [12] Timeout               
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1: PCIe Bus Error: 
severity=Correctable, type=Data Link Layer, (Transmitter ID)
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1:   device [1002:aae0] 
error status/mask=00001000/00002000
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1:    [12] Timeout               
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER: Correctable error 
message received from 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER: found no error details 
for 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER: Multiple Correctable 
error message received from 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER: found no error details 
for 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER: Multiple Correctable 
error message received from 0000:00:02.0
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: PCIe Bus Error: 
severity=Correctable, type=Data Link Layer, (Transmitter ID)
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0:   device [8086:2f04] error 
status/mask=00001040/00002000
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0:    [ 6] BadTLP                
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0:    [12] Timeout               
Oct 10 20:46:37 xev kernel: pcieport 0000:00:02.0: AER:   Error of this Agent 
is reported first
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0: PCIe Bus Error: 
severity=Correctable, type=Data Link Layer, (Transmitter ID)
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0:   device [1002:6987] error 
status/mask=00001100/00002000
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0:    [ 8] Rollover              
Oct 10 20:46:37 xev kernel: amdgpu 0000:02:00.0:    [12] Timeout               
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1: PCIe Bus Error: 
severity=Correctable, type=Data Link Layer, (Transmitter ID)
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1:   device [1002:aae0] 
error status/mask=00001100/00002000
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1:    [ 8] Rollover              
Oct 10 20:46:37 xev kernel: snd_hda_intel 0000:02:00.1:    [12] Timeout        

On that system I took the CPU out and reinstalled it with new heatsink paste on the theory that it might not have made good contact with some of the pins. The system also has one DIMM slot not working which can be a symptom of poor seating of the CPU. Doing that made no difference to the DIMM slot (I had bought the system for $50 in “unknown condition”) but the video has worked correctly since. It has been suggested to me that reseating the CPU didn’t directly affect the issue and that just taking the system apart could have addressed an issue of the GPU not making good contact in the PCIe slot.

It has been suggested that I could try “contact cleaner” which can be obtained from automotive supply stores among other places. I’m hesitant to put that in a PCIe slot but putting it on the connector of the card and then polishing it off seems like something to consider. Another suggestion was to use isopropyl alcohol to wash the contacts. I guess washing a PCIe slot out with isopropyl alcohol and leaving it for hours to dry is an option as a last resort.

For the moment it seems to be fine but I am not certain that the problem is gone forever. At the moment my main aim is to have these systems keep working until after the release of DDR6 workstations which is when I expect DDR5 workstations to become affordable on all the second hand sites.

• [1] https://etbe.coker.com.au/2025/04/05/hp-z840/
• [2] https://etbe.coker.com.au/2025/06/20/intel-arc-b580-pcie-slot-size/
• [3] https://etbe.coker.com.au/2025/04/05/hp-ml110-gen9-z640/

For at least 12 years laptops have been defaulting to not having the traditional PC 101 key keyboard function key functionality and instead have had other functions like controlling the volume and have had a key labelled Fn to toggle the functions. It’s been a BIOS option to control whether traditional function keys or controls for volume etc are the default and for at least 12 years I’ve configured all my laptops to have the traditional function keys as the default.

Recently I’ve been working in corporate IT and having exposure to many laptops with the default BIOS settings for those keys to change volume etc and no reasonable option for addressing it. This has made me reconsider the options for configuring these things.

Here’s a page listing the standard uses of function keys [1]. Here is a summary of the relevant part of that page:

• F1 key launches help doesn’t seem to get much use. The main help option in practice is Google (I anticipate controversy about this and welcome comments) and all the software vendors are investigating LLM options for help which probably won’t involve F1.
• F2 is for renaming files but doesn’t get much use. Probably most people who use graphical file managers use the right mouse button for it. I use it when sorting a selection of photos.
• F3 is for launching a search (which is CTRL-F in most programs).
• ALT-F4 is for closing a window which gets some use, although for me the windows I close are web browsers (via CTRL-W) and terminals (via CTRL-D).
• F5 is for reloading a page which is used a lot in web browsers.
• F6 moves the input focus to the URL field of a web browser.
• F8 is for moving a file which in the degenerate case covers the rename functionality of F2.
• F11 is for full-screen mode in browsers which is sometimes handy.

The keys F1, F3, F4, F7, F9, F10, and F12 don’t get much use for me and for the people I observe. The F2 and F8 keys aren’t useful in most programs, F6 is only really used in web browsers – but the web browser counts as “most programs” nowadays.

Here’s the description of Thinkpad Fn keys [2]. I use Thinkpads for fun and Dell laptops for work, so it would be nice if they both worked in similar ways but of course they don’t. Dell doesn’t document how their Fn keys are laid out, but the relevant bit is that F1 to F4 are the same as on Thinkpads which is convenient as they are the ones that are likely to be commonly used and needed in a hurry.

I have used the KDE settings on my Thinkpad to map the function F1 to F3 keys to the Fn equivalents which are F1 to mute-audio, F2 for vol-down, and F3 for vol-up to allow using them without holding down the Fn key while having other function keys such as F5 and F6 have their usual GUI functionality. Now I have to could train myself to use F8 in situations where I usually use F2, at least when using a laptop.

The only other Fn combinations I use are F5 and F6 for controlling screen brightness, but that’s not something I use much.

It’s annoying that the laptop manufacturers forced me to this. Having a Fn key to get extra functions and not need 101+ keys on a laptop size device is a reasonable design choice. But they could have done away with the PrintScreen key to make space for something else. Also for Thinkpads a touch pad is something that could obviously be removed to gain some extra space as the Trackpoint does all that’s needed in that regard.

• [1] https://tinyurl.com/22x67h6z
• [2] https://tinyurl.com/26sj5eoo

There are many negative articles about “AI” (which is not about actual Artificial Intelligence also known as “AGI”). Which I think are mostly overblown and often ridiculous.

Resource Usage

Complaints about resource usage are common, training Llama 3.1 could apparently produce as much pollution as “10,000 round trips by car between Los Angeles and New York City”. That’s not great but when you compare to the actual number of people doing such drives in the US and the number of people taking commercial flights on that route it doesn’t seem like such a big deal. Apparently commercial passenger jets cause CO2 emissions per passenger about equal to a car with 2 people. Why is it relevant whether pollution comes from running servers, driving cars, or steel mills? Why not just tax polluters for the damage they do and let the market sort it out? People in the US make a big deal about not being communist, so why not have a capitalist solution, make it more expensive to do undesirable things and let the market sort it out?

ML systems are a less bad use of compute resources than Bitcoin, at least ML systems give some useful results while Bitcoin has nothing good going for it.

The Dot-Com Comparison

People often complain about the apparent impossibility of “AI” companies doing what investors think they will do. But this isn’t anything new, that all happened before with the “dot com boom”. I’m not the first person to make this comparison, The Daily WTF (a high quality site about IT mistakes) has an interesting article making this comparison [1]. But my conclusions are quite different.

The result of that was a lot of Internet companies going bankrupt, the investors in those companies losing money, and other companies then bought up their assets and made profitable companies. The cheap Internet we now have was built on the hardware from bankrupt companies which was sold for far less than the manufacture price. That allowed it to scale up from modem speeds to ADSL without the users paying enough to cover the purchase of the infrastructure. In the early 2000s I worked for two major Dutch ISPs that went bankrupt (not my fault) and one of them continued operations in the identical manner after having the stock price go to zero (I didn’t get to witness what happened with the other one). As far as I’m aware random Dutch citizens and residents didn’t suffer from this and employees just got jobs elsewhere.

There are good things being done with ML systems and when companies like OpenAI go bankrupt other companies will buy the hardware and do good things.

NVidia isn’t ever going to have the future sales that would justify a market capitalisation of almost 4 Trillion US dollars. This market cap can support paying for new research and purchasing rights to patented technology in a similar way to the high stock price of Google supported buying YouTube, DoubleClick, and Motorola Mobility which are the keys to Google’s profits now.

The Real Upsides of ML

Until recently I worked for a company that used ML systems to analyse drivers for signs of fatigue, distraction, or other inappropriate things (smoking which is illegal in China, using a mobile phone, etc). That work was directly aimed at saving human lives with a significant secondary aim of saving wear on vehicles (in the mining industry drowsy drivers damage truck tires and that’s a huge business expense).

There are many applications of ML in medical research such as recognising cancer cells in tissue samples.

There are many less important uses for ML systems, such as recognising different types of pastries to correctly bill bakery customers – technology that was apparently repurposed for recognising cancer cells.

The ability to recognise objects in photos is useful. It can be used for people who want to learn about random objects they see and could be used for helping young children learn about their environment. It also has some potential for assistance for visually impaired people, it wouldn’t be good for safety critical systems (don’t cross a road because a ML system says there are no cars coming) but could be useful for identifying objects (is this a lemon or a lime). The Humane AI pin had some real potential to do good things but there wasn’t a suitable business model [2], I think that someone will develop similar technology in a useful way eventually.

Even without trying to do what the Humane AI Pin attempted, there are many ways for ML based systems to assist phone and PC use.

ML systems allow analysing large quantities of data and giving information that may be correct. When used by a human who knows how to recognise good answers this can be an efficient way of solving problems. I personally have solved many computer problems with the help of LLM systems while skipping over many results that were obviously wrong to me. I believe that any expert in any field that is covered in the LLM input data could find some benefits from getting suggestions from an LLM. It won’t necessarily allow them to solve problems that they couldn’t solve without it but it can provide them with a set of obviously wrong answers mixed in with some useful tips about where to look for the right answers.

Jobs and Politics

Noema Magazine has an insightful article about how “AI” can allow different models of work which can enlarge the middle class [3].

I don’t think it’s reasonable to expect ML systems to make as much impact on society as the industrial revolution, and the agricultural revolutions which took society from more than 90% farm workers to less than 5%. That doesn’t mean everything will be fine but it is something that can seem OK after the changes have happened. I’m not saying “apart from the death and destruction everything will be good”, the death and destruction are optional. Improvements in manufacturing and farming didn’t have to involve poverty and death for many people, improvements to agriculture didn’t have to involve overcrowding and death from disease. This was an issue of political decisions that were made.

The Real Problems of ML

Political decisions that are being made now have the aim of making the rich even richer and leaving more people in poverty and in many cases dying due to being unable to afford healthcare. The ML systems that aim to facilitate such things haven’t been as successful as evil people have hoped but it will happen and we need appropriate legislation if we aren’t going to have revolutions.

There are documented cases of suicide being inspired by Chat GPT systems [4]. There have been people inspired towards murder by ChatGPT systems but AFAIK no-one has actually succeeded in such a crime yet. There are serious issues that need to be addressed with the technology and with legal constraints about how people may use it. It’s interesting to consider the possible uses of ChatGPT systems for providing suggestions to a psychologist, maybe ChatGPT systems could be used to alleviate mental health problems.

The cases of LLM systems being used for cheating on assignments etc isn’t a real issue. People have been cheating on assignments since organised education was invented.

There is a real problem of ML systems based on biased input data that issue decisions that are the average of the bigotry of the people who provided input. That isn’t going to be worse than the current situation of bigoted humans making decisions based on hate and preconceptions but it will be more insidious. It is possible to search for that so for example a bank could test it’s mortgage approval ML system by changing one factor at a time (name, gender, age, address, etc) and see if it changes the answer. If it turns out that the ML system is biased on names then the input data could have names removed. If it turns out to be biased about address then there could be weights put in to oppose that.

For a long time there has been excessive trust in computers. Computers aren’t magic they just do maths really fast and implement choices based on the work of programmers – who have all the failings of other humans. Excessive trust in a rule based system is less risky than excessive trust in a ML system where no-one really knows why it makes the decisions it makes.

Self driving cars kill people, this is the truth that Tesla stock holders don’t want people to know.

Companies that try to automate everything with “AI” are going to be in for some nasty surprises. Getting computers to do everything that humans do in any job is going to be a large portion of an actual intelligent computer which if it is achieved will raise an entirely different set of problems.

I’ve previously blogged about ML Security [5]. I don’t think this will be any worse than all the other computer security problems in the long term, although it will be more insidious.

How Will It Go?

Companies spending billions of dollars without firm plans for how to make money are going to go bankrupt no matter what business they are in. Companies like Google and Microsoft can waste some billions of dollars on AI Chat systems and still keep going as successful businesses. Companies like OpenAI that do nothing other than such chat systems won’t go well. But their assets can be used by new companies when sold at less than 10% the purchase price.

Companies like NVidia that have high stock prices based on the supposed ongoing growth in use of their hardware will have their stock prices crash. But the new technology they develop will be used by other people for other purposes. If hospitals can get cheap diagnostic ML systems because of unreasonable investment into “AI” then that could be a win for humanity.

Companies that bet their entire business on AI even when it’s not necessarily their core business (as Tesla has done with self driving) will have their stock price crash dramatically at a minimum and have the possibility of bankruptcy. Having Tesla go bankrupt is definitely better than having people try to use them as self driving cars.

• [1] https://thedailywtf.com/articles/ai-the-bad-the-worse-and-the-ugly
• [2] https://etbe.coker.com.au/2024/04/26/humane-ai-pin/
• [3] https://www.noemamag.com/how-ai-could-help-rebuild-the-middle-class/
• [4] https://tinyurl.com/2b95xdzs
• [5] https://etbe.coker.com.au/2025/05/30/machine-learning-security/

To run a SMP system with multiple CPUs you need to have CPUs that are “identical”, the question is what does “identical” mean. In this case I’m interested in Intel CPUs because SMP motherboards and server systems for Intel CPUs are readily available and affordable. There are people selling matched pairs of CPUs on ebay which tend to be more expensive than randomly buying 2 of the same CPU model, so if you can identify 2 CPUs that are “identical” which are sold separately then you can save some money. Also if you own a two CPU system with only one CPU installed then buying a second CPU to match the first is cheaper and easier than buying two more CPUs and removing a perfectly working CPU.

Intel (R) Xeon (R)
E5-2640V4
SR2NZ 2.40GHZ
J717B324 (e4)
7758S4100843

Above is a pic of 2 E5-2640v4 CPUs that were in a SMP system I purchased along with a plain ASCII representation of the text on one of them. The bottom code (starting with “77”) is apparently the serial number, one of the two codes above it is what determines how “identical” those CPUs are.

The code on the same line as the nominal clock speed (in this case SR2NZ) is the “spec number” which is sometimes referred to as “sspec” [1].

The line below the sspec and above the serial number has J717B324 which doesn’t have a google hit. I looked at more than 20 pics of E5-2640v4 CPUs on ebay, they all had the code SR2NZ but had different numbers on the line below. I conclude that the number on the line below probably indicates the model AND stepping while SR2NZ just means E5-2640v4 regardless of stepping. As I wasn’t able to find another CPU on ebay with the same number on the line below the sspec I believe that it will be unreasonably difficult to get a match for an existing CPU.

For the purpose of matching CPUs I believe that if the line above the serial number matches then the CPUs can be used together. I am not certain that CPUs with this number slightly mismatching won’t work but I definitely wouldn’t want to spend money on CPUs with this number being different.

smpboot: CPU0: Intel(R) Xeon(R) CPU E5-2699A v4 @ 2.40GHz (family: 0x6, model: 0x4f, stepping: 0x1)

When you boot Linux the kernel identifies the CPU in a manner like the above, the combination of family and model seem to map to one spec number. The combination of family, model, and stepping should be all that’s required to have them work together.

I think that Intel did the wrong thing in not making this clearer. It would have been very easy to print the stepping on the CPU case next to the sspec or the CPU model name. It also wouldn’t have been too hard to make the CPU provide the magic number that is apparently the required match for SMP to the OS. Having the Intel web site provide a mapping of those numbers to steppings of CPUs also shouldn’t be difficult for them.

If anyone knows more about these issues please let me know.

• [1] https://tinyurl.com/28nedh9r

I’ve been looking at computer hardware on AliExpress a lot recently and I saw an advert for a motherboard which can take 256G DDR4 RDIMMs (presumably LRDIMMs). Most web pages about DDR4 state that 128G is the largest possible. The Wikipedia page for DDR4 doesn’t state that 128G is the maximum but does have 128G as the largest size mentioned on the page.

Recently I’ve been buying 32G DDR4 RDIMMs for between $25 and $30 each. A friend can get me 64G modules for about $70 at the lowest price. If I hadn’t already bought a heap of 32G modules I’d buy some 64G modules right now at that price as it’s worth paying 40% extra to allow better options for future expansion.

Apparently the going rate for 128G modules is $300 each which is within the range for a hobbyist who has a real need for RAM. 256G modules are around $1200 each which is starting to get a big expensive. But at that price I could buy 2TB of RAM for $9600 and the computer containing it still wouldn’t be the most expensive computer I’ve bought – the laptop that cost $5800 in 1998 takes that honour when inflation is taken into account.

DDR5 RDIMMs are currently around $10/GB compared to DDR4 for $1/GB for 32G modules and DDR3 for $0.50/GB. DDR6 is supposed to be released late this year or early next year so hopefully enterprise grade systems with DDR5 RAM and DDR5 RDIMMs will be getting cheaper on ebay by the end of next year.

It seems to me that we haven’t had much change in the overall design of desktop PCs since floppy drives were removed, and modern PCs still have bays the size of 5.25″ floppy drives despite having nothing modern that can fit in such spaces other than DVD drives (which aren’t really modern) and carriers for 4*2.5″ drives both of which most people don’t use. We had the PC System Design Guide [1] which was last updated in 2001 which should have been updated more recently to address some of these issues, the thing that most people will find familiar in that standard is the colours for audio ports. Microsoft developed the Legacy Free PC [2] concept which was a good one. There’s a lot of things that could be added to the list of legacy stuff to avoid, TPM 1.2, 5.25″ drive bays, inefficient PSUs, hardware that doesn’t sleep when idle or which prevents the CPU from sleeping, VGA and DVI ports, ethernet slower than 2.5Gbit, and video that doesn’t include HDMI 2.1 or DisplayPort 2.1 for 8K support. There are recently released high-end PCs on sale right now with 1gbit ethernet as standard and hardly any PCs support resolutions above 4K properly.

Here are some of the things that I think should be in a modern PC System Design Guide.

Power Supply

The power supply is a core part of the computer and it’s central location dictates the layout of the rest of the PC. GaN PSUs are more power efficient and therefore require less cooling. A 400W USB power supply is about 1/4 the size of a standard PC PSU and doesn’t have a cooling fan. A new PC standard should include less space for the PSU except for systems with multiple CPUs or that are designed for multiple GPUs.

A Dell T630 server has an option of a 1600W PSU that is 20*8.5*4cm = 680cc. The typical dimensions of an ATX PSU are 15*8.6*14cm = 1806cc. The SFX (small form factor variant of ATX) PSU is 12.5*6.3*10cm = 787cc. There is a reason for the ATX and SFX PSUs having a much worse ratio of power to size and that is the airflow. Server class systems are designed for good airflow and can efficiently cool the PSU with less space and they are also designed for uses where people are less concerned about fan noise. But the 680cc used for a 1600W Dell server PSU that predates GaN technology could be used for a modern GaN PSU that supplies the ~600W needed for a modern PC while being quiet. There are several different smaller size PSUs for name-brand PCs (where compatibility with other systems isn’t needed) that have been around for ~20 years but there hasn’t been a standard so all white-box PC systems have had really large PSUs.

PCs need USB-C PD ports that can charge a laptop etc. There are phones that can draw 80W for fast charging and it’s not unreasonable to expect a PC to be able to charge a phone at it’s maximum speed.

GPUs should have USB-C alternate mode output and support full USB functionality over the cable as well as PD that can power the monitor. Having a monitor with a separate PSU, a HDMI or DP cable to the PC, and a USB cable between PC and monitor is an annoyance. There should be one cable between PC and monitor and then keyboard, mouse, etc should connect to the monior.

All devices that are connected to a PC should use USB-C for power connection. That includes monitors that are using HDMI or DisplayPort for video, desktop switches, home Wifi APs, printers, and speakers (even when using line-in for the audio signal). The European Commission Common Charger Directive is really good but it only covers portable devices, keyboards, and mice.

Motherboard Features

Latest verions of Wifi and Bluetooth on the motherboard (this is becoming a standard feature).

On motherboard video that supports 8K resolution. An option of a PCIe GPU is a good thing to have but it would be nice if the motherboard had enough video capabilities to satisfy most users. There are several options for video that have a higher resolution than 4K and making things just work at 8K means that there will be less e-waste in future.

ECC RAM should be a standard feature on all motherboards, having a single bit error cause a system crash is a MS-DOS thing, we need to move past that.

There should be built in hardware for monitoring the system status that is better than BIOS beeps on boot. Lenovo laptops have a feature for having the BIOS play a tune on a serious error with an Android app to decode the meaning of the tune, we could have a standard for this. For desktop PCs there should be a standard for LCD status displays similar to the ones on servers, this would be cheap if everyone did it.

Case Features

The way the Framework Laptop can be expanded with modules is really good [3]. There should be something similar for PC cases. While you can buy USB devices for these things they are messy and risk getting knocked out of their sockets when moving cables around. While the Framework laptop expansion cards are much more expensive than other devices with similar functions that are aimed at a mass market if there was a standard for PCs then the devices to fit them would become cheap.

The PC System Design Guide specifies colors for ports (which is good) but not the feel of them. While some ports like Ethernet ports allow someone to feel which way the connector should go it isn’t possible to easily feel which way a HDMI or DisplayPort connector should go. It would be good if there was a standard that required plastic spikes on one side or some other way of feeling which way a connector should go.

GPU Placement

In modern systems it’s fairly common to have a high heatsink on the CPU with a fan to blow air in at the front and out the back of the PC. The GPU (which often dissipates twice as much heat as the CPU) has fans blowing air in sideways and not out the back. This gives some sort of compromise between poor cooling and excessive noise. What we need is to have air blown directly through a GPU heatsink and out of the case. One option for a tower case that needs minimal changes is to have the PCIe slot nearest the bottom of the case used for the GPU and have a grille in the bottom to allow air to go out, the case could have feet to keep it a few cm above the floor or desk. Another possibility is to have a PCIe slot parallel to the rear surface of the case (right angles to the other PCIe slots).

A common case with desktop PCs is to have the GPU use more than half the total power of the PC. The placement of the GPU shouldn’t be an afterthought, it should be central to the design.

Is a PCIe card even a good way of installing a GPU? Could we have a standard GPU socket on the motherboard next to the CPU socket and use the same type of heatsink and fan for GPU and CPU?

External Cooling

There are a range of aftermarket cooling devices for laptops that push cool air in the bottom or suck it out the side. We need to have similar options for desktop PCs. I think it would be ideal to have a standard attachments for airflow on the front and back of tower PCs. The larger a fan is the slower it can spin to give the same airflow and therefore the less noise it will produce. Instead of just relying on 10cm fans at the front and back of a PC to push air in and suck it out you could have a conical rubber duct connected to a 30cm diameter fan. That would allow quieter fans to do most of the work in pushing air through the PC and also allow the hot air to be directed somewhere suitable. When doing computer work in summer it’s not great to have a PC sending 300+W of waste heat into the room you are in. If it could be directed out a window that would be good.

Noise

For restricting noise of PCs we have industrial relations legislation that seems to basically require that workers not be exposed to noise louder than a blender, so if a PC is quieter than that then it’s OK. For name brand PCs there are specs about how much noise is produced but there are usually caveats like “under typical load” or “with a typical feature set” that excuse them from liability if the noise is louder than expected. It doesn’t seem possible for someone to own a PC, determine that the noise from it is what is acceptable, and then buy another that is close to the same.

We need regulations about this, and the EU seems the best jurisdiction for it as they cover the purchase of a lot of computer equipment that is also sold without change in other countries. The regulations need to also cover updates, for example I have a Dell T630 which is unreasonably loud and Dell support doesn’t have much incentive to be particularly helpful about it. BIOS updates routinely tweak things like fan speeds without the developers having an incentive to keep it as quiet as it was when it was sold.

What Else?

Please comment about other things you think should be standard PC features.

• [1] https://en.wikipedia.org/wiki/PC_System_Design_Guide
• [2] https://en.wikipedia.org/wiki/Legacy-free_PC
• [3] https://frame.work/au/en/marketplace/expansion-cards

I recently had someone describe a Mac Mini as a “workstation”, which I strongly disagree with. The Wikipedia page for Workstation [1] says that it’s a type of computer designed for scientific or technical use, for a single user, and would commonly run a multi-user OS.

The Mac Mini runs a multi-user OS and is designed for a single user. The issue is whether it is for “scientific or technical use”. A Mac Mini is a nice little graphical system which could be used for CAD and other engineering work. But I believe that the low capabilities of the system and lack of expansion options make it less of a workstation.

The latest versions of the Mac Mini (to be officially launched next week) have up to 64G of RAM and up to 8T of storage. That is quite decent compute power for a small device. For comparison the HP ML 110 Gen9 workstation I’m currently using was released in 2021 and has 256G of RAM and has 4 * 3.5″ SAS bays so I could easily put a few 4TB NVMe devices and some hard drives larger than 10TB. The HP Z640 workstation I have was released in 2014 and has 128G of RAM and 4*2.5″ SATA drive bays and 2*3.5″ SATA drive bays. Previously I had a Dell PowerEdge T320 which was released in 2012 and had 96G of RAM and 8*3.5″ SAS bays.

In CPU and GPU power the recent Mac Minis will compare well to my latest workstations. But they compare poorly to workstations from as much as 12 years ago for RAM and storage. Which is more important depends on the task, if you have to do calculations on 80G of data with lots of scans through the entire data set then a system with 64G of RAM will perform very poorly and a system with 96G and a CPU less than half as fast will perform better. A Dell PowerEdge T320 from 2012 fully loaded with 192G of RAM will outperform a modern Mac Mini on many tasks due to this and the T420 supported up to 384G.

Another issue is generic expansion options. I expect a workstation to have a number of PCIe slots free for GPUs and other devices. The T320 I used to use had a PCIe power cable for a power hungry GPU and I think all the T320 and T420 models with high power PSUs supported that.

I think that a usable definition of a “workstation” is a system having a feature set that is typical of servers (ECC RAM, lots of storage for RAID, maybe hot-swap storage devices, maybe redundant PSUs, and lots of expansion options) while also being suitable for running on a desktop or under a desk. The Mac Mini is nice for running on a desk but that’s the only workstation criteria it fits. I think that ECC RAM should be a mandatory criteria and any system without it isn’t a workstation. That excludes most Apple hardware. The Mac Mini is more of a thin-client than a workstation.

My main workstation with ECC RAM could run 3 VMs that each have more RAM than the largest Mac Mini that will be sold next week.

If 32G of non-ECC RAM is considered enough for a “workstation” then you could get an Android phone that counts as a workstation – and it will probably cost less than a Mac Mini.

• [1] https://en.wikipedia.org/wiki/Workstation

On the 18th of May I blogged about my new 5120*2160 monitor [1]. One thing I noted was that one Netflix movie had run in an aspect ratio that used all space on the monitor. I still don’t know if the movie in question was cropped in a letterbox manner but other Netflix shows in “full screen” mode don’t extend to both edges. Also one movie I downloaded as in 3840*1608 resolution which is almost exactly the same aspect ratio as my monitor. I wonder if some company is using 5120*2160 screens for TVs, 4K and FullHD are rumoured to be cheaper than most other resolutions partly due to TV technology being used for monitors. There is the Anamorphic Format of between 2.35:1 and 2.40:1 [2] which is a close match for the 2.37:1 of my new monitor.

I tried out the HDMI audio on a Dell laptop and my Thinkpad Yoga Gen3 and found it to be of poor quality, it seemed limited to 2560*1440, at this time I’m not sure how much of the fault is due to the laptops and how much is due to the monitor. The monitor docs state that it needs HDMI version 2.1 which was released in 2017 and my Thinkpad Yoga Gen3 was released in 2018 so probably doesn’t have that. The HDMI cable in question did 4K resolution on my previous monitor so it should all work at a minimum of 4K resolution.

The switching between inputs is a problem. If I switch between DisplayPort for my PC and HDMI for a laptop the monitor will usually timeout before the laptop establishes a connection and then switch back to the DisplayPort input. So far I have had to physically disconnect the input source I don’t want to use. The DisplayPort switch that I’ve used doesn’t seem designed to work with resolutions higher than 4K.

I’ve bought a new USB-C dock which is described as doing 8K which means that as my Thinkpad is described as supporting 5120×2880@60Hz over USB-C I should be able to get 5120*2160 without any problems, however for unknown reasons I only get 4K. For work I’m using a Dell Latitude 7400 2in1 that’s apparently only capable of 4096*2304 @24 Hz which is less pixels than 5120*2160 and it will also only do 4K resolution. But for both those cases it’s still a significant improvement over 2560*1440. I tested with a Dell Latitude 7440 which gave the full 5120*2160 resolution, I was unable to find specs on what the maximum resolution of the 7440 is. I also have bought DisplayPort switch rated at 8K resolution. I got a switch that doesn’t also do USB because the ones that do 8K resolution and USB are about $70. The only KVM switch I saw for 8K resolution at a reasonable price was one designed for switching between two laptops and there doesn’t seem to be any adaptors to convert from regular DisplayPort to USB-C alternative mode so that wasn’t viable. Currently I have the old KVM switch used for USB only (for keyboard and mouse) and the new switch which only does DisplayPort. So I have two buttons to push when switching between input sources which isn’t too bad.

It seems that for PCs resolutions with more pixels than 4K are as difficult and inconvenient now as 4K was 6 years ago when I started doing it. If you want higher than 4K resolution to just work at this time then you need Apple hardware.

The monitor has a series of modes for different types of output, I’ve found “standard” to be good for text and “movie” to be good for watching movies/TV and for playing RTS games. I previously wrote about how to use ddcutil to use a monitor as a KVM switch [3], unfortunately I can’t do this with the new monitor as the time that the monitor waits for a good signal on a new input after changing is shorter than the time taken for Linux on the laptops I’m using to enable HDMI output. I’ve found the following commands to do the basics.

# get display mode
ddcutil getvcp DC
# set standard mode
ddcutil setvcp DC 0
# set movie mode
ddcutil setvcp DC 03

Now that I have that going the next thing I want to do is to have it switch between “standard” and “movie” modes when I switch keyboard focus.

• [1] https://etbe.coker.com.au/2024/05/18/kogan-5120×2160-40-monitor/
• [2] https://en.wikipedia.org/wiki/Anamorphic_format
• [3] https://etbe.coker.com.au/2022/07/19/ddc-switch/

I wrote a comment on a social media post where someone claimed that there’s no computer advances in the last 12 years which got long so it’s worth a blog post.

In the last decade or so new laptops have become cheaper than new desktop PCs. USB-C has taken over for phones and for laptop charging so all recent laptops support USB-C docks and monitors with USB-C docks built in have become common. 4K monitors have become cheap and common and higher than 4K is cheap for some use cases such as ultra wide. 4K TVs are cheap and TVs with built-in Android computers for playing internet content are now standard. For most use cases spinning media hard drives are obsolete, SSDs large enough for all the content most people need to store are cheap. We have gone from gigabit Ethernet being expensive to 2.5 gigabit being cheap.

12 years ago smart phones were very limited and every couple of years there would be significant improvements. Since about 2018 phones have been capable of doing most things most people want. 5yo Android phones can run the latest apps and take high quality pics. Any phone that supports VoLTE will be good for another 5+ years if it has security support. Phones without security support still work and are quite usable apart from being insecure. Google and Samsung have significantly increased their minimum security support for their phones and the GKI project from Google makes it easier for smaller vendors to give longer security support. There are a variety of open Android projects like LineageOS which give longer security support on a variety of phones. If you deliberately choose a phone that is likely to be well supported by projects like LineageOS (which pretty much means just Pixel phones) then you can expect to be able to actually use it when it is 10 years old. Compare this to the Samsung Galaxy S3 released in 2012 which was a massive improvement over the original Galaxy S (the S2 felt closer to the S than the S3). The Samsung Galaxy S4 released in 2013 was one of the first phones to have FullHD resolution which is high enough that most people can’t easily recognise the benefits of higher resolution. It wasn’t until 2015 that phones with 4G of RAM became common which is enough that for most phone use it’s adequate today.

Now that 16G of RAM is affordable in laptops running more secure OSs like Qubes is viable for more people. Even without Qubes, OS security has been improving a lot with better compiler features, new languages like Rust, and changes to software design and testing. Containers are being used more but we still aren’t getting all the benefits of that. TPM has become usable in the last few years and we are only starting to take advantage of what it can offer.

In 2012 BTRFS was still at an early stage of development and not many people wanted to use it in production, I was using it in production then and while I didn’t lose any data from bugs I did have some downtime because of BTRFS issues. Now BTRFS is quite solid for server use.

DDR4 was released in 2014 and gave significant improvements over DDR3 for performance and capacity. My home workstation now has 256G of DDR4 which wasn’t particularly expensive while the previous biggest system I owned had 96G of DDR3 RAM. Now DDR5 is available to again increase performance and size while also making DDR4 cheap on the second hand market.

This isn’t a comprehensive list of all advances in the computer industry over the last 12 years or so, it’s just some things that seem particularly noteworthy to me.

Please comment about what you think are the most noteworthy advances I didn’t mention.