The latest version of Intel Arc GPU Graphics Software introduced an interesting change that isn’t reflected in the Release Notes. The installer of the 101.4578 beta drivers add a “Compute Improvement Program” (CIP) component as part of the “typical” setup option that is enabled by default. Under the “custom” installer option that you have to activate manually, you get to select which components to install. The Compute Improvement Program can be unchecked here, to ensure data collection is disabled. The benignly named CIP is a data collection component that tracks your PC usage and performance in the background (not just that of the GPU), so Intel can use the data to improve its future products. Intel created a dedicated webpage that spells out what CIP is, and what its scope of data collection is; where is says that CIP “does not collect your name, email address, phone number, sensitive personal information, or physical location (except for country).”
NVIDIA’s and AMD’s drivers also contain telemetry collection software, and only AMD tries to be as transparent as possible about it by offering a check box during installation, whereas Intel and NVIDIA hide it behind the “custom” option.
Needless to say, Linux users don’t have to worry about this.
Everyone should read the intel link and expand the topics. It’s worse than the quoted text claims…
Intel monitoring website usage…really? I find this kind of shocking, our web usage has no bearing whatsoever on intel arc drivers. Intel generates tracking IDs for each computer (ie cookies), giving them the means to track users over time without any other identifying information, which they claim not to collect…
Whenever these privacy invading topics come up, I always want to see actual log entries to see exactly what they are recording. I think this is a completely reasonable request, unfortunately though services rarely ever provide or disclose exactly what they are taking. And there may be other ways users can be correlated with other databases. For example while intel says the computing improvement program doesn’t log IPs, it doesn’t strictly mean that intel aren’t using internet firewalls that do.
I’m not alleging intel have malicious intent, but if they don’t really need the data to improve ARC hardware, and it doesn’t appear that they need all these categories, then it’s better for users that they simply don’t collect it at all. The thing that people may overlook is that this data can be subpoenaed and used in lawsuits or even in criminal cases. In other words, the fact that intel doesn’t collect “personal information” doesn’t negate that the ownership of IDs can be tracked and proven after the fact.
Intel needs to learn about https://martinfowler.com/bliki/Datensparsamkeit.html
ssokolow,
Thanks for the link!
I’ll try to remember this word “Datensparsamkeit”, it’s definitely a useful concept!
Alfman,
I think your context is different, but Intel needing this data is also reasonable.
Yet, whenever someone does need to collect data, they tend to go overboard: “let’s collect everything today, we might find a use tomorrow”.
Building graphics drivers is hard, especially when client programs are buggy (most of the “game optimized” driver work is actually fixing the bugs by game developers by detecting their patterns at runtime, and replacing them with better commands dynamically). So, Intel really needs what application is currently running, how is the frame rate, memory throughput, API overhead, and such.
But why do they need a list of websites we access?
sukru,
I don’t follow what you mean? I believe users should be entitled to see exactly what data gets submitted regardless of context.
That’s the problem though, isn’t it? I mean, web data has zero relevance to what intel needs to know for ARC drivers. Intel are doing it to spy on users because they can, not because they need it to improve ARC drivers. I don’t even know how one would play devil’s advocate to pretend it’s to benefit ARC users, it really isn’t.
Sure, I’m not against data collection tools, but I still think they should be explicitly submitted by owners on a need to know basis and not turned into 24/7 spyware.
This sort of nonsense is why the “I just want things to work” gaming PC I recently built from hand-me-down or spare parts currently runs Windows 7 rather than Windows 10, and has been hooked up such that its network connection can only be used to exchange files with my main PC, not the Internet.
(Anything too old to Just Work™ on 64-bit Windows 7 gets installed on the Windows XP or Windows 98SE/DOS machines in my retro-hobby corner, where it can also Just Work™.)
ssokolow,
I concur, and have kept windows 7 around for as long as a could. But realistically my experience was getting worse and worse on account of microsoft intentionally sabotaging standard software runtime dependencies to break on windows 7. The vast majority of developers compile with default settings, which means that any recent software builds stop running on windows 7 even though the APIs are 100% compatible. *Grumble*
This is always so funny to me. It’s EASIER to get (especially old) Windows things to work on Linux these days than it often is on modern Windows (especially games), and they often run a lot better. On top of that, Linux or even macOS are FAR more reliable than Windows even pretends to be, and you don’t have to constantly spend 4 hours updating everything.
If you want something that “just works” – why are you still on Windows? My guess is it’s a sunk cost fallacy. Most people don’t consider how much special knowledge they have built up around how to make things work on Window, compared with how much special knowledge they’d have to procure to make things work on Linux. Linux is actually easier, and also more reliable. It’s just that you haven’t put in the time the same way you have with Windows. The reliable part is the important part. With Windows you have to KEEP investing that time and effort – like all the time, because it’s so unreliable. It’s a total waste of effort, and you don’t even get to own the software. Windows will never be “yours”.
I’m not. I’ve been daily-driving Linux since the early 2000s when I got fed up with Windows XP.
I have a “Just Works” Windows 7 machine because my daily driver Linux machine…
1. Is multitasking its ass off and gamemoderun isn’t enough.
2. Is an Athlon II X2 270 from 2011 where I had to apply every performance trick I could think of just to get A Hat in Time to be playable because its CPU is right on the edge of the system requirements before adding DX→Vulkan translation overhead to the mix.
3. Struggles to game and play YouTube videos at the same time, while I can comfortably do that if I just run YouTube and the games on separate PCs.
4. Has its SATA ports off the end of the PCI-E x16 slot and is full to the brim with hard drives, so I can’t install something bulkier than the GeForce GTX750 I’ve been running for nearly a decade because I only discovered just now that Amazon.ca apparently doesn’t see Silverstone Tek low-profile angled SATA cables as discontinued the way other sources like NewEgg.ca do.
5. Runs 24/7 in my bedroom, which means I’m reluctant to install a GPU that might be noisier.
By contrast, my gaming PC…
1. Is a hand-me-down HP prebuilt of a similar vintage with an Intel Core i3 that’s 1.4 times as fast, but has a motherboard that can’t take more than 8GiB of RAM, so I can’t use it to replace my main PC. (My main PC has 32GiB.)
2. Has my spare 650W SeaSonic PSU and the hand-me-down Radeon HD 5870 from a friend in it… because HP never released a BIOS update to make it compatible with the more capable GeForce GTX760 my brother upgraded off of.
3. Can dedicate its full resources to the game with no multitasking.
4. Doesn’t leave me debugging things like “When did Binding of Isaac: Rebirth start crashing on startup under the Wine version I pinned for it in PlayOnLinux and how do I fix it?”
5. Doesn’t require a very specific Wine version to keep Dungeon Keeper 2 from crashing.
6. Allows me to fullscreen Dungeons of Dredmor without having to source or make a non-Steam-distributed build of Proton to keep fullscreening from changing my desktop video mode.
7. Allows me to hook two mice into two PCs and be able to play a mouse-capturing game on my center monitor (KVM switched) AND control YouTube on my left monitor.
8. Plays 99% of the games I care about perfectly, despite being made of tech old enough for people to give it to me as hand-me-downs. (There are a couple of GOG games like Skyrim and Itorah that it’s not good enough for, but I’ve got such a giant backlog that I don’t care.)
Don’t think of it as “I don’t run Linux”. Think of it as “I have a Linux PC and spent $45 CAD on a DVI/USB KVM switch to comfortably share its central monitor with a ‘gaming console’, but without the console part”.
ssokolow,
Agreed. While linux gaming has gotten much better over the years thanks to more games being ported and better windows OS emulation. it’s still not as easy as using the OS the games were designed for. Some people find linux good enough for them, but windows is better supported by most game studios.
I would think that you might be able to pull this off if your script-foo is strong enough 🙂 Rather than using an emulated mouse, couldn’t you configure KVM to forward a USB mouse into the VM?
This hasn’t been my experience, at least for older games (Binding of Isaac is new enough to count as “not old”). Windows can be considerably worse at running older games than Wine/Linux/Proton. Of course, I’ve also had the opposite problem. I suppose it depends on what you want to run.
CaptainN-,
Fair enough. I’ve had numerous issues trying windows games on linux distros (don’t mean to blow this out of proportion, but just saying I understand where the critics are coming from). In terms of windows, I haven’t been running it on any personal PCs since win7 anyways, so I can’t say what gaming is like on win8 and newer. Surely it’s still got to be better than linux, right? I’d be really shocked if linux has better compatibility than windows with old games, that doesn’t seem likely to me. But I’m open to being convinced if there’s data that makes a strong case for it.
Keyboard/Video/Mouse switch, not Kernel Virtual Machine.
A literal physical piece of hardware that functions as a DVI/USB switchbox and a USB hub and switches a DVI port and two USB ports between my Linux daily driver and my Windows 7 gaming machine when I double-tab Scroll Lock.
I have two mice because:
1. If I switch inputs using the monitor’s OSD and use Barrier (a software KVM switch forked from Synergy), games with mouselook are unable to capture the mouse. (i.e. You’re turning right in a first-person or over-the-shoulder third-person 3D game and, suddenly, your mouse is on the PC like it would be if you were running an ordinary windowed desktop app.)
…plus, Barrier isn’t privileged to synthesize input while UAC prompts are open.
2. My KVM switch and my preferred mouse for my daily driver PC (a Logitech G203 mouse with five buttons and a laser sensor) don’t get along. (It’s prone to going non-responsive about a minute after toggling the switch)
That’s why I have both the hardware KVM switch and an HDMI line bypassing it and Barrier set up. I can either have a normal KVM switch experience, with the center monitor, keyboard, and mouse toggling between both PCs, or use the monitor’s OSD to switch only the monitor to the gaming machine while leaving the keyboard and mice on Linux. (Useful if the only input the game needs is my Xbox 360 pad and I want control a YouTube video or scroll through a strategy guide on the side monitors.)
Either way, it’s nice to have such a hard separation between the two machines, since nothing that happens on the Linux machine can cause the game to stutter or forced-AltTab out of it or anything like that, and, when the keyboard and mouse are KVMed to the Linux side, nothing the game does can lock my mouse out of manipulating PDFs or websites or whatever and double-tapping Scroll Lock to bring the keyboard and mouse back and forth can’t mess with games that react badly to Alt+Tab.
ssokolow (Hey, OSNews U2F/WebAuthn is broken on Firefox!) ,
I sure misinterpreted that!
I’m not familiar with that software KVM switch, but a physical KVM switch shouldn’t have this problem.
That’s no good. I’m sure there are KVM switches that work. Of course the problem is there are probably hundreds of clones from the same chinese OEM source and as a consumer you have no idea which ones are the good ones. I’ve been there (with different hardware).
Yeah, whatever fits your needs! I have a no-name switch between my computers that has a manual toggle button and no keyboard shortcuts. It additionally came with a remote control, I don’t need to switch on a regular basis, but I have it because I hate switching wires out when I need to work on another machine.
I also use a lantronix spider. Make no mistake, it’s a low quality feed, but it’s worked out well for connecting to headless equipment at client facilities. Otherwise I’d have to lug around a monitor/keyboard/mouse. There’s an open source ip KVM that I wanted to try, but everything based on RPI has been chronically out of stock.
CaptainN-,
Agree that familiarity is a large part of it. It’s not just linux that can feel unfamiliar either, I use macos so intermittently that I’ve struggled to do some pretty basic things like installing applications that are blocked by macos security. On the one hand, it is legitimate criticism, but on the other a knowledgeable macos user would have memorized the steps to take. When you’re a daily user, things become second nature.
In my view normal users can use linux assuming it’s preinstalled and users stick to the repos for software. That’s fairly easy to do. Alas, linux often gets a bad rap on PC because the majority of users are installing it themselves without any official manufacturer support, which gets them into trouble. This is a higher technical barrier to entry than other operating systems tend to face because they already came bundled. I find it unfortunate that aside from chromebooks, more PC manufacturers don’t officially support linux as an option. Manufacturers are not going to do it because they have no financial incentive, not to mention pressure by microsoft to only ship windows.
https://www.osnews.com/story/136392/the-only-pc-ever-shipped-with-beos-preinstalled/
But naturally, if more systems dual booted alternatives, a lot more users would naturally become familiar with other operating systems.
I do recall on Linus Tech Tips, Linus tried a “Linux challenge” and ran in to numerous issues that stemmed from this choice of distro. First Pop!_OS, which ships with 2 flavors, regular and nVidia – and the nVidia one DOES have issues with compatibility, which he was unable to get passed. But then he went with I think it was Arch Linux, which is NOT something I’d recommend to a newbie. He’d have been more successful, IMHO, if he went with Ubuntu, or Mint. But that’s part of the problem with how people talk about Linux – Linux is not an operating system the same way Windows is. Pop!_OS, Ubuntu, Arch, etc. – those are EACH comparable to Window, but they are all very distinct from each other.
I agree with your comments on the lack of manufacturer support. I’ve had pretty good luck with home grown towers, and on Lenovo laptops, but I can imagine some other vendors machines being kind of finicky.
CaptainN-,
Yeah no kidding, an advanced linux distro was really a poor choice for a newbie, Surely if he asked around he would have gotten a very different recommendation, which implies he either didn’t ask around, or he ignored the recommendations. I’d go further and say that ideally they should be using linux preinstalled on officially supported hardware so that they don’t have to take responsibility for supporting untested hardware themselves. Preinstalling linux gives them the same entry point as they would get with a windows or macos PC. Otherwise it’s kind of cheeky to give those operating systems the benefit of a different starting line.
Yes indeed.
Most x86 systems I’ve tried linux on work, but of course those looking to criticize linux can point to those that don’t work as a justification for linux being worthless, haha.
Honestly I’ve gotten linux to boot on almost 100% of the hardware I’ve tried. The problem areas, if there are any, are usually with peripherals or unaccellerated graphics. For example my newest motherhboard’s network card wasn’t supported and I had to use a PCI card.
Often the motherboard’s temp/fan/voltage/etc sensor chips aren’t supported (most users don’t look at this stuff anyways, but I do like to be able to monitor temps from the OS).
In one case I had a laptop where UEFI secure boot was hardcoded to microsoft’s keys, and it gave me lots of trouble. I couldn’t run my own distro, I resigned myself to running ubuntu on it since they are signed under microsoft’s keys.
https://wiki.ubuntu.com/UEFI/SecureBoot
Needless to say I vehemently oppose this, owners should not need microsoft’s permission to run other operating systems. This seems to be fairly rare, but we need to keep our guard up though because a growing dependency on microsoft secure boot keys is extremely dangerous for linux and alt-os.
As a Linux user, this is why I would prefer Linux stay a nerd thing, like ad blocking used to be. Once it gets popular, caused perhaps by reaching critical mass with gamers, the shady behavior of the software industry will come right over, too. Each program you run will want to collect telemetry, run in the background constantly when you’re not using it, change your homepage, and switch your default web browser!
Seriously, if you don’t think it can happen to Linux, look at Android.
And I’m still not sold on the concept of telemetry in software. A lot of us really want Intel to succeed in the GPU market. Intel doesn’t have to go far to find what customers want; faster game performance and support in more rendering engines like Iray. Intel has been doing amazing stuff with Blender, but it will take a lot of work to catch up with Nvidia across the board. If you want customer feedback, just have a forum and a bug reporting system.
Many distros will always be for nerds. Just because most people use Android, doesn’t mean Gentoo isn’t nerdy.
kbd,
I think you’re right, intel are doing this for intel’s benefit, not for users, which is unfortunate. That said, I think there is value in having a tool for on-demand submission, like crash reports and at the request of tech support. We wouldn’t be here criticizing intel for that. Beyond that though intel are clearly crossing the privacy line unnecessarily. Monitoring application usage without the user taking action is cringey, Monitoring web usage…WTF intel?! Even on a granular level this is indefensible.
> Needless to say, Linux users don’t have to worry about this.
Unless said users are using one of google’s linux variants.