Recent content by The Historical Fidelity

My first reaction was “are you serious Microsoft?”

Wrong, and editing my comment with your disinformation is very inappropriate. Your “research article” is filled with marketing fluff and hopeful thinking, I wouldn’t trust anything beyond the actual experimental data they do provide. Everything else is classified as “forward-facing statements”...

Read that article again. They do not already have “EUV Lasers figured out”. Their laser’s conversion efficiency is ~50% worse than a similar laser system built in 2007 at the university of Florida, and their pulse frequency is far below what is required and is the hardest thing to increase due...

Yes, the 2070 super is close, but less close versus the 1080 that matches the 5050 in both number of cores and identical bandwidth, whereas the 2070 super is only a match in core number.

Nvidia’s official 5050 specs states it is using 20 Gbps GDDR6 on a 128-bit bus which does indeed equal the same bandwidth as 10 Gbps GDDR5X on a 256-bit bus, whereas TPU’s 5050 specs page is most likely a holdover from using the 3050 page as a template to create the new 5050 page since it states...

To be honest, the 5050 is almost identical to the GTX 1080 besides the added RTX logic units on the die. Both have 2,560 cores, 8 GB of memory, and the same memory bandwidth. The only difference is in how each card reaches the same bandwidth, IE: gtx 1080 has a 256-bit bus running 10 Gbps GDDR5X...

All due respect, but I believe you are also missing the point. I am not a “believer” or an “fan”, and I didn’t “believe the spin” or was “misled” at all. I do not understand the level of assumption you had to apply in order to formulate your reply. Thus I will lay my point out as clear as...

Appreciate it.

You are skirting around my point yet again. It’s a SIMPLE point. If you say you are going for significant marketshare (to summarize Frank Azor: specifically to the point that game designers see AMD’s platform as important enough to optimize for in the PC segment), then you produce the hell out...

I’m sorry, but no, companies have entire divisions dedicated to predicting market size. AMD clearly was not looking to gain market-share, they hedged their bets and allocated a conservative amount of wafers because they didn’t think they would be as competitive as they turned out to be. This is...

I just don’t get it…A few months before the 9000 series launch, the head of graphics at AMD emphasized that this generation they are focusing on gaining market share. And then they launch it without the required volume to actually do what they say they want to achieve??? This is not a good look...

Since HBM is itself a stacked DRAM solution, the only difference with Intel’s offering will the efficiency of routing data and power lines. So my question is, will this “efficient wiring” be internal to the stack only? So that the package is interchangeable with HBM on physical boards? Does this...

I read somewhere that HarmonyOS was a “fake it till you make it” OS in that they heavily relied on Android and slowly replaced the android code with their own replacement code over time. Kind of similar to a competitor reverse engineering their IP. So in essence, yes, Huawei can claim they have...

I highly doubt the company’s assertions…

Can anyone confirm if the memory bus is running at 8 Gbps? Because adding 8 Gbps capable memory chips is very different from the memory controller being able to run the chips at that speed. Most likely the memory controller is still running at 7 Gbps