Recent content by Filiprino

And Microsoft pretends game streaming is OK. ROFL.

This is great.

Zen 2 + Navi = WIN.

Combat, combat and more combat.

I still have the OEM Windows 95 CD-ROM and Microsoft Works 4.0 for Windows CD-ROM.

Windows 10 abuses the user by pushing things he does not want. Even with batch files to disable telemetry there are still things hanging around like web search integrated into the desktop even with Cortana turned off, which implies sending data packets to Bing servers when you are searching the...

Ownership of game saves, mods, latency, image quality, performance (saturated servers), spying on clients, ... too much negatives.

I would expect 6 cores per CCX. The 10-15% IPC improvement is wonderful. It will put AMD at latest gen Intel level or better. For those who ask, more cores come handy for software based multimedia decoding. Not relying in hardware devices is good. Problems arise when you use hardware video...

¿Linux? GNU is the OS. Linux is a kernel. Usually when you use GNU with Linux is spelled GNU/Linux or GNU+Linux or GNU-Linux to give credit where it is due. Now that we know what we are talking about, Snaps occupy too much space. It is better if you do not depend on Snaps. Snaps should be the...

Bo-ring.

Timeless classic. Instant buy. Now I need an HDR screen :( My GFX card is a Polaris so I think I am ready on the computer side.

Lock-free Linux kernel exists since a long time ago. Windows NT still struggles with kernel locks slowing down systems with high core counts. You can design an algorithm and your data structures to access data without using locks. Coordination can be made only between relevant groups of cores...

Amdah'ls Law graphs depending on parallelizable fraction: https://askldjd.com/2010/01/30/some-plots-on-amdahls-law/ Algorithms are usually 90%+ parallelizable. Typical HPC workloads are 99% parallel. And people try to design algorithms to have 99% parallel fractions. On top of that there's...

The only way to continue improving is adding more cores. And making them power efficient. Which means that you will end up with multiple small in-order cores and a few big out-of-order cores. Then you have to use runtime libraries to exploit those types of chips. But then you get into commercial...

Just keep supporting GNU/Linux ecosystem please. Wayland is taking a long time to come around argh.