OK, Eegore, what proportion of AMD do you use and why do you use it instead of buying Intel Inside?

OK, Eegore, what proportion of AMD do you use and why do you use it instead of buying Intel Inside?


 I couldn't tell you exactly the percentage of AMD vs Intel I have purchased in my lifetime.  If I buy a block of HP units and they have AMD processors then I use AMD because it is inside the HP computers I purchased.  Typically availability of items and if they can perform the task(s) requested is how the decision is made.

A dozen groups have just reported that AMD's stock price has just lapped over Intel's falling stock prices ---- the worm has turned .......





This is another "crossing of the graph lines" event ......   stock prices are now crossed over each other and are moving very fast in their new directions.

Also note that lots of folks are yelling for Bob Swan's head for having mis-managed Intel so very very badly ......

"Do you specify memory amounts or any other features when buying?"

 I do not, I only purchase what is dictated by the project that is being worked on so I would say the requirements of the project specify.


Have you ever gotten something that wouldn't do the jobs you needed once you got it in hand?

 Not from a hardware standpoint that I am aware of.  I have had software that wasn't compatible with certain programs etc.

https://www.fool.com/investing/2020/07/24/why-intel-stock-plunged-today.aspx

OK, this is a Motley Fool staff writer "explaining" the Intel stock debacle that took place yesterday.   He does a fair job of it, but misses a few points about tweeked Intel benchmarks and Intel's lying BS performance ratings.

AMD isn't giving Intel an inch

Digital video game distribution platform Steam releases a hardware and software survey every month. According to the platform's latest monthly report for May, the number of Steam users with an AMD processor increased to 22.45%. This was a slight bump from April's 21.71%.

While Intel commands the rest of the market, it is worth noting that Steam's monthly surveys indicate that AMD's share increased in four out of the first five months of the year. Of course, the survey is limited, as it only captures data from Steam users. Still, the platform had 95 million monthly active users last year, so it gives us some insight into the wider gaming CPU market.

What's more, this isn't the only piece of data that suggests AMD's growing clout in the CPU market. German retailer Mindfactory's sales data for the week of June 1 to June 7 reveals that it sold 5,270 AMD processors compared to just 770 Intel processors. Again, this data has limitations, but it does hint at a worrying trend for Intel.

Chipzilla launched its Comet Lake desktop processors toward the end of this past April. Tom's Hardware had even called the i5-10600K processor a "mainstream gaming champ" thanks to its strong single-threaded performance. Its special price of $262 during the launch added to its attractiveness and helped with the sell through numbers as well.

But the launch of these new chips hasn't altered the emerging trend as Mindfactory data reveals that AMD's Ryzen 5 3600 processor remains the best seller. The Core i5-10600K moved just 40 units. AMD's latest update to its Ryzen lineup means that Intel's new chips will now face a bigger challenge as AMD could reduce the prices of their older chips to clear inventory space for the new chips that are coming in the next few weeks.

That's because AMD's latest chips deliver a minimum improvement of 4%-10% per iteration as compared to their existing versions. The updated Ryzen 5 3600XT carries a sticker price of $249, as does the Ryzen 5 3600X. If AMD reduces the price of their existing comparable chips as they have been doing during the ramp-downs (clearing out some replacement shelf-space) recently then the AMD price-to-performance ratio could improve further and make it even more difficult for Intel's Comet Lake processors to make a dent in the market.

Turning on the heat

AMD has also made it clear that the launch of its upcoming Zen 3 processors is on track for 2020. This is more bad news for Intel. With this move, AMD is likely to further refine its 7-nanometer (nm) manufacturing process that powers the current generation of Ryzen chips. Intel's Comet Lake processors, on the other hand, are based on a 14nm process, as the chip giant has so far failed to make the jump to 10nm desktop processors successfully.

The rumor mill suggests that Intel's 10nm Alder Lake offerings that could potentially compete with AMD's 7nm process will only partially begin to arrive next year (and now will likely be delayed well past that). If that is the case, Intel runs the risk of remaining well behind AMD on the technology curve, as Intel doesn't expect to launch any 5nm processors until 2022-2023 per its current product roadmap.

So far, AMD's process node advantage has helped it deliver competitive processors and attractive prices, which have eaten into Intel's market share. The AMD product roadmap ahead suggests that AMD will keep its foot on the accelerator and remain the top semiconductor stock going forward.

https://seekingalpha.com/article/4338843-taiwan-semiconductor-manufacturing-c...

This is a finance magazine Intel fan boy searching for something good to say about Intel's far far future plans.   Mind you, he does not understand the delays just put forth by Intel will be ADDING YEARS to the timelines he reports, he is just regurgitating what Intel has released in print for "Intel's far future plans".

And it is obvious someone from Intel has shared "the true meanings" of Intel's 10nm and 7nm because he quotes that propaganda as fact.

Still, taken in perspective it gives you timelines that will only get worse (with more delay) as Intel fumbles the ball time after time trying to do things that TSMC cannot pull off the very first time out of the gate ......

The precept of this article should be that Intel will buy (or rent) current TSMC style lithography going forward.    Intel has said OFFICIALLY that they are going to do this, but Intel has certainly not ordered any new lithography scanners to put into TSMC to put any reality behind those empty words.

Not having ordered 2-3 year lead time ASML scanner lines several years ago means Intel has to rent 5nm production time from TSMC/Apple  or Samsung.   Samsung doesn't HAVE enough equipment to handle Intel's volumes, so forget Samsung as a source for more than a single chipset item or two.

Intel trying to rent enough huge massive amounts 5nm production time from TSMC/Apple is a joke right now too as all TSMC/Apple's 5nm production is booked up for up to 3 years ahead at this stage of things.   Apple paid for the original TSMC 5nm lines and Apple will not turn them loose until they are up and running on 3nm with their full current Apple production.   So, 3nm will be ramping up for world-wise use before there is enough free 5nm capacity to even start to handle Intel's needed volumes.  

Furthermore, the bottom graph that is showing Intel making advances to lithography stages well before TSMC does kinda loses track of the fact that Intel is having TSMC build all their advanced lithography for them going forward.   This last graph is pure optimistic Intel fanboy fiction from ages past, in other words.

This article writer totally dismisses all the effects of 14 layer deep direct burn EUV  lithography and what that means for large amounts of very fast on chip multi-level cache memory that AMD is currently using to their advantage.   Intel knows nothing about this layered deep burn trick and it is not reflected in any of their current chipset designs.  So to say that Intel just picks it up right along with gate all around chip construction techniques is just plain silly talk.

Intel has lost Jim Heller as a resource (he quit in frustration) and they have only Renchalla left now as their tech leader.  Renchalla, the man who insisted that Intel go hire Keller, the guy who taught Renchalla how to do it when talking about fine feature advanced lithography.





This very last graph is pure "Intel wishes it were so" and it also assumes that there is copious amounts of free TSMC production space to plug Intel into ..... and we know that this is not the case.

OR ...... it assumes Intel can actually do 5nm followed by 3nm gate all around all in house at will, which is another joke to beat all jokes.




THIS LAST GRAPH IS MOSTLY BOGUS NONSENSE

So, add in a 2 year gap between the graph lines and remove all the line overlaps ---- this is closer to best case reality.  

Or, simply being completely realistic, there should only be the TSMC graph line showing in the years past 2020 .....

And what is with the weird years with quarters shown on the X axis of this last graph --- this is all complete Intel Fanboy BS that was written a long long time ago, isn't it?

https://wccftech.com/intel-ponte-vecchio-gpu-tsmc-6nm/

Intel Ponte Vecchio GPU Will Be Made On TSMC’s 6nm Process, CPUs Could Be Made On TSMC 3nm As Well

A huge report from DigiTimes has leaked out claiming that Intel's Ponte Vecchio GPU has found its home on TSMC's lower-cost 6nm process. The leak originally sprung from a paywalled DigiTimes article and has since been corroborated by China Times as well in the last few hours. The implication of Intel shifting major products to TSMC is huge and the report even indicates that talks to move Intel's bread and butter CPUs to the pure-play foundry are being held as well.

Intel's Ponte Vecchio GPU will be made on the TSMC 6nm process, books 180,000 orders in advance
Intel's 10nm process was supposedly to be actually better than TSMC's 7nm process (MTr/mm2) and it was roughly comparable to TSMC's 6nm process. This means products that are fabricated on TSMC's 6nm process (which is an optimized and lower cost version of 7nm) will have roughly the same density as a 10nm Intel part.

This footprint thing is still problematic in the sense because Intel's 7nm was supposed to have even higher MTr/mm2 and could impact the socket footprint of projects dependant on Ponte Vecchio such as the Aurora supercomputer.    New main processor sockets on the way, boys?

This thing reads like a blow by blow of next year's AMD vs Intel battle.   Intel's ability to be in the fight at all depends on Intel booking enough of TSMC's 6nm and 3nm capacity to build what they need to be build.

AMD may well have booked first, however, and AMD will likely come out at least a quarter to a half year earlier, since they planned to roll down to 6nm TSMC long ago (back AMD was a key partner in the development of TSMC's 6nm process stream) and this was back when Intel was going it alone on their 10nm and 7nm stuff.

https://www.tomshardware.com/news/tsmc-2nm-process-development-cpus-5nm-proce...

https://www.tomshardware.com/news/tsmc-fab-3nm-5nm-process-intel-samsung


TSMC Starts 2nm Process Development for Fast, Efficient Chips




At the start of this year we heard that TSMC was investing heavily in extra 5nm fabrication buildings, and it's no secret that the next step after that is TSMC's 3nm process as TSMC has just finished building the first 3nm plant buildings and the 3nm campus proper.   3nm multi-level 3-D direct burn EUV production line equipment are being installed and multi-level prove out runs are going on as we speak.

The Taiwanese silicon manufacturer is now looking beyond that 3nm process too, and TSMC has just announced to shareholders that it is starting full technical development of its 2nm lithographic process with key customers, as was spotted by DigiTimes.

Painstakingly few details are available now about the 2nm process. All we know is that TSMC is starting development -- though it's safe to assume the end product will be very fast and more efficient than anything on the market today.

Currently, TSMC's 7nm process is at its mature peak, receiving huge numbers of orders from AMD for its Ryzen 3000-series CPUs and Navi graphics cards.  TSMC's 6nm is planned out as the logical extension of TSMC 7nm, an extension that can run off the same existing 7nm EUV production equipment after an upgrade/rebuild.

On the 5nm front, TSMC is working with brand new ASML 14 layer direct burn EUV lithography scanner lines, similar to what Samsung is accomplishing. The two chipmakers are neck-in-neck in the 5nm silicon design race at this time. According to DigiTimes, TSMC is expecting 10% of this year's revenue to come from its many layer 5nm EUV lines.

With the news last year that TMSC had been cleared by their Board of Directors to begin working on the 3nm production site, the cost was estimated at $19.5 billion and production slated to start in late 2022 or early 2023.  Earlier this year, C.C. Wei, CEO of TSMC, said that 3nm development was “going well” and that it was engaging with early customers on the latter stages of full technology definition.

So, 3nm is getting real now and is installing the first 3nm lines, 5nm is completely in the hole for first large Apple production runs which are happening as we speak and 2nm is buying land and engaging early customers on the early states of 2nm technology definition.



Meanwhile, Intel just placed their first 180,000 shortfall chipset order at TSMC for their "complete production failure" shortfalls in Intel 10nm chipsets .....