66cats

Not sure what the Chalkboard benchmark tests, but MotionMark results seem a bit more reflective of reality. Why else would HW acceleration be a part of all modern browsers?

But the results don't change, nearly identical to switching "Use HW acceleration when available": #enable-gpu-rasterization disabled: #enable-gpu-rasterization enabled:

Results make a bit more sense in this one. Win 7 with GPU: Win 7 w/o GPU: XP (w/o GPU):

Mine (both on the same box). XP: 7: Unexpected (Speedometer 2/3 scores better on 7) 7 with "Use graphics acceleration when available" off: P.S. Not sure what that benchmark tests, but scrolling without HW acceleration feels choppy (~20fps) on this box, smooth/fluid (screen refresh rate) with HW acceleration on.

Oh, got you. Never used one,

Extracting the executable to a folder & running chrome.exe for the first time still creates new registry entries & directories (so it won't be portable). These won't affect your Thorium (or any other browser) installs, if that's an issue.

Looks like the conversation was locked by win32, the repository owner. Guessing because it was a simple statement, not a conversation opener. If you still wish to chat about it, all you need do is click a button.

Just a guess, but nowhere. The typically cached elements are simply re-downloaded, again and again, as if the page was visited for the first time. If @Kmuland's internet bandwidth is greater than HDD bandwidth, disabling the cache makes sense. Again, just guessing, no idea how browsers work.

Try the 32-bit build of Thorium, see if (with Chrome-xp-api-adapter) it's quicker than the 64-bit build. Oddly enough, for me, it is (on x64 SP2). Works on my machine™ (Thorium/XP x64)

Same here. Might have something to do with @AstroSkipper's single-core rig, happens to me all the time posting from a Coppermine PIII (on any forum). BTW, IDA-RE's dll does seem to speed up Thorium/Supermium (launching for the first time, and at least as far as Speedometer 2 is concerned). Curious factoids: it's now called "chrome-xp-api-adapter.dll" instead of "progwrp.dll" (i renamed it), and it's 24KB vs. 131KB of the original.

The same logic that makes Supermium closed source (if any part of the project is not open source, it's not open source) should make Linux closed source (its kernel includes [proprietary] binary blobs). I guess those are the only open source Linux distros OTOH, most distros (including all the popular ones, the ones where WiFi & most HW works, ones using the standard, unmodded linux kernel) are proprietary. At least by that logic :\ Something tells me reliance on leaked/reversed MS code is the reason no source has been published for that .dll. Again, could be dead wrong.

Please make sure to take a couple of seconds to perform a simple search before posting. Which part of the quoted text do you disagree with? Be concise. I concede that, like Linux, Supermium contains closed source code. If you feel this makes those projects proprietary, i got nothing.

Supermium source, and [at least] the binary of that .dll is, and always was, publicly available. Not sure how not publishing the source would affect people who wanted to support win32/needed an installer.

So that was the one that got famous? (forgive me, i'm a bit out of the loop when it comes to browser starter apps)

Probably the same people who started spreading rumors that Linux is open source. I'm having trouble reconciling how something shared with a small circle of friends could be thought of as famous. Horse-famous, maybe?

Huh? I'm asking Dixel if he's sure the source wasn't published (and not getting an answer). Also pointing out the question is not rhetorical &, not being a coder, i wouldn't know where to look. Not sure how i could've been more explicit. having trouble reconciling this with

Are you sure win32 didn't publish the source? Thought he did. (real question, not a coder & wouldn't know where to look) You didn't publish the source? How come?

Goes something like: 1. win32 publishes progwrp.dll code, which is picked up by Alex313031, and, apparently, IDA-RE 2. IDA-RE, who is not win32, releases a version of win32's dll, calls it progwrp.dll v. 1.2.0.5035 3. IDA-RE, who is not win32, takes down his work (progwrp.dll v. 1.2.0.5035) -- both the source and the binaries ∴ win32's Supermium is not open source, QED. ?

Using Thorium_SSE4_122.0.6261.168_WINXP_x64.zip for XP x64, Vista & 7 (multiboot, so literally the same copy). Will try Thorium_AVX2_123.0.6312.133.zip in Vista 64 & post results in a few. UPDATE: tried AVX2 (this box is 4770k), AVX & even SSE3 flavors on W7, all from https://github.com/Alex313031/Thorium-Win/releases. Each one gives me the "not a valid win32 app" error. AVX2 flavor performs roughly the same on W10: 144, 156, 145 The weird thing is on W10, the XP/legacy version (Thorium_SSE4_122.0.6261.168_WINXP_x64.zip) benches better (196, 200, 204) than the AVX2 version meant for Win10. In Vista extended kernel (Nov. 2022), i get Supermium on Vista extended kernel (Nov. 2022): 199, 203, 200 Thorium (XP SSE4) on Vista extended kernel: 192, 197, 198

Doesn't seem to. Just retested three times with just one tab: 136, 125, 131. Then again, not really a night-and-day difference between the two, can't notice it outside benchmarks.

I think that was it, thanks.

Is for me too (in XP), both about the same in Vista & 7. BTW, sorry for linking images, can't embed them (images of any size) for some reason.

Could force a 9xx driver (install from the device manager, 'let me choose,' etc.) & get 2D acceleration. Worked for me with 1050 & 1070 (also on 8th gen B chipset). Plug the display into the motherboard FTW?

Works for me, though drops frames in XP, as do earlier versions. Works fine on Vista+ (decent HW, e.g. browsers like 360Chrome & Mypal don't drop frames).

Does MS actually have a secret XP ESU program, or are you talking about some company selling alleged security updates? Always assumed XP was down to to running aging signage, ancient-but-cool lab gear & crusty CNC rigs, that sort of thing, air-gapped from reality.