If you’re having trouble installing or updating to Comodo Dragon 36, check your CPU. If it does not support the SSE2 instruction set, you’re probably stuck using Dragon 33. Google Chrome stopped supporting non-SSE2 CPUs with version 35, and it seems reasonable that Dragon 36 followed suit.
I don’t know if this is a source code issue, a compiler issue, or both. Apparently, the latest version of Microsoft’s compiler defaults to SSE2 support without SSE support. If it is only a compiler issue, then I humbly request that the Dragon team include SSE support in their builds. If that adds too much bloat, then an SSE only build would be appreciated.
I recommend analyzing your CPU because some P4 CPUs don’t support SSE2 while some Pentium M CPUs do (e.g. the one on my trusty Dell Latitude D610 running Win7). I had a hunch that the CPU was the problem on a spare box I use mostly for scanning. I used CPU-Z (v1.71.1) from CPUID.com to analyze its CPU, and found that the AMD Athlon XP 1500+ supported SSE, but not SSE2.
Thanks to Comodo for providing an alternative to Big G’s browser.
Thank you for your reply and for the link to “Issue 349320: SSE2 on Windows” on code.google.com. I had seen this page and that’s how I knew about the compiler issue. Since I’m not a programming jock, I stopped with the first comment.
Upon further reading I found comment 51:
The ship has sailed on dynamic SSE detection; M35 is built to require SSE2 throughout, even on Windows (see comment 38 and earlier). So you might as well force it to on in skia too. [sic]
So, Chromium’s base code (> M34) requires SSE2. Therefore, Comodo Dragon (> 33) also requires SSE2. Fair enough.
The system requirements for Dragon should be updated to reflect this fact.
My understanding of Issue 349320 is that Chromium without SSE2 has no future. While it might be possible to compile Chromium without SSE2, the result will probably not be useable, “as other parts of chromium may either require SSE2 or unconditionally use SSE2 instructions”¹.
While I understand and respect your wish to keep using your hardware, I think it is reasonable to require support for an instruction-set which has been available since 2001. SSE2 (and higher) is becoming increasingly important for browsers’ performance. One reason is that audio and video is increasingly being handled by the browser (rather than plugins). Look, for example, at libvpx, a library for encoding and decoding video (VP8 and VP9), which is included in Chromium (and Firefox). Here are the latest SSE2-changes: chromium Git repositories - Git at Google You can also search for SSSE3 and SSE4.1.
I’m convinced that Firefox will eventually require SSE2 too, it’s just unknown when it will happen.