Alphane_Moon

Valve has confirmed that gamers anticipating a successor to the Steam Deck will need to wait until more compelling processors hit the market.

Not to mention a measure of stabilization in the tariff environment.

I am honestly surprised that IBM decided to go with "AI" as a major feature of their new mainframe product. In a sense they an excellent market of legacy customers who are willing to pay almost any price for continued iterative updates and high class professional support.

Let's hope they can get beyond 60% and to 80%. I am not saying this because I have Samsung stock (I actually have Intel stock), but we really need a viable competitor to TSMC.

I have heard of some of these CPUs over the years.

An interesting one for me was the "Intel Xeon LV 1.6 D1 (Prestonia)":

The Prestonia core was basically the Pentium 4 Northwood with SMP (symmetric multiprocessing) and HyperThreading added as standard features. With the sub-$200 1.6GHz Xeon drawing a frugal 1.274v, overclockers generally couldn't take advantage of voltage headroom as most boards were voltage-locked. However, simply raising the FSB would net 2.6GHz.

For the more adventurous, three hard mods could yield a 100% overclock (or more!): the U-Wire mod which involved bridging two (1.5v) or three (1.6v) sets of socket pins, the BSEL mod to isolate or break CPU pins and raise the FSB limit to 200MHz, and the vDIMM mod to raise RAM voltage.

Those willing to push the limits of the technology could be rewarded with a 3.2GHz dual processor performance king for around $700 (CPUs, coolers, board, and RAM).

Even without the hard mods, this is a solid CPU with basic FSB raise considering the price.

Getting in the GPU game in the current environment is a huge undertaking. Arc was clearly a first generation product. Battlemage is far more viable, if you're going for a relatively cheap dGPU, there is no better option than B580/B670.

According to Bloomberg, the U.S. relies on imports from China for a majority of the following products by value: game consoles (86%, $6 billion), PC monitors (79%, $5 billion), smartphones (73%, $41 billion), lithium-ion batteries (70%, $16 billion), and laptops (66%, $32 billion).

Not to mention countries other than China (particularly ones that are involved in tech hardware manufacturing like Vietnam and Thailand) are also under tariffs.

Sounds like a solution in search of a problem.

Not to mention that the Android XR platform will be shut down pretty soon (this is Google after all).

There are numerous ways to measure AI throughput, making it difficult to compare chips. Google is using FP8 precision as its benchmark for the new TPU, but it's comparing it to some systems, like the El Capitan supercomputer, that don't support FP8 in hardware. So you should take its claim that Ironwood "pods" are 24 times faster than comparable segments of the world's most powerful supercomputer with a grain of salt.

This is not a grain of salt. This is premeditated lying.

Honestly the whole article reminds of this:

https://www.youtube.com/watch?v=GFRzIOna2oQ

Can you configure the output for something other than a podcast discussion and have a more dry, technical lecture format?

I have always done copy/paste for password confirmations. "Yes" and "confirm" is new for me, I haven't heard of such inputs. 🤣🥃

I guess the next step would be solving the issue. Recognizing that you may need to take inspiration from other countries and also apply a more realistic risk management/tolerance approach (the matters we are discussions often include difficult choices).

I am talking about hypotheticals. Just like saying Apple will tolerate 2% profit margin vs their current ~35%.

It's a thought experiment.