Quote from: Non IT Expert on November 03, 2023, 01:09:01This is very impressive CPU wise, if not as revolutionary as the original M1 chip, even with the delays, no matter how you dice it. But then again not surprising considering the team it comes from, so we were kind of expecting this. I mean think about it, you're getting Zen 4 class performance or above but with apple level efficiency

Such has not been true before; we must await independent benchmarks (not: Geekbench) to see if it is true now. Also Apple cannot beat physics. In a small form factor, it is silent and not that fast - or fast and not that silent.

QuoteBigger concern for me is the GPU side. [...] there doesn't seem any options for those who were looking for console / M-max class perf alternative.

The option is: x64 + dGPU.

QuoteWould be very interesting to see this chip in a device paired a Nvidia dGPU

M3 + Nvidia dGPU suppose volatile storage use can be sorted out? You are dreaming. :)

Dunno why people are focusing on the CPU side. This is very impressive CPU wise, if not as revolutionary as the original M1 chip, even with the delays, no matter how you dice it. But then again not surprising considering the team it comes from, so we were kind of expecting this. I mean think about it, you're getting Zen 4 class performance or above but with apple level efficiency, without even having to resort to Big.little core design (that both Intel / Apple use, and AMD soon will too). That's insane!

Bigger concern for me is the GPU side. It seems roughly competitive with MTL/Strix but that's not saying much since both those have relatively small die area for the iGPU. Now that Strix Halo seems DOA (massive delays), there doesn't seem any options for those who were looking for console / M-max class perf alternative.

Also, people have mentioned that what will make or break this product is how good the x86 to arm translation is in upcoming win 12. This comment may age poorly but based on what I see so far, it doesn't seem Microsoft is actually concerned at all or even trying to make a major push / improvement in this regard (Cuz otherwise we would of heard it by now in leaks/rumors as an upcoming major breakthrough). They seem hellbent on pushing the AI narrative hype through all their services even if there really no end use case for it for the avg consumer, which is a shame.

Lastly, there is the pricing. Knowing Qcomm and their recent SoC price increases (SD8G2 / SD8G3), this won't be cheap. This will probably be in devices that are similar in price range to Apple (so like 2000-3000 range). As such I'll probably just get a switch 2 instead and wait for it to get hacked, then run win 12 on it. Yeh, it'll have far slower and less CPU cores. But it'll be only $400ish and have close to similar GPU perf.

Would be very interesting to see this chip in a device paired a Nvidia dGPU (or even integrated Tegra like the Geforce ION days), but not holding my breath for anything as don't believe these two companies are on the best of terms.

The M3 only beats the Snapdragon X Elite in Single core in Geekbench 6.2 when running on Windows. When running the Linux version the scores are reversed and the M3 looses. This is probably a better comparison as MacOS and Linux are POSIX like operating systems.

Despite any efforts and tricks with TDP in PL2 mode, Intel always beats AMD in single-threaded performance in the same class, taking into account real availability on the market.

By the way, AMD here published a preliminary forecast for revenue - growth in sales of processors for desktops and laptops in the 3rd quarter in unit terms by more than 60% (but revenue growth only by 42%, due to a fall in the average price of processors) -
ir.amd.com/sec-filings/content/0000002488-23-000195/amd-20230930.htm?TB_iframe=true&height=auto&width=auto&preload=false

AMD points out that most of the growth came from laptop processors.

At the same time, Intel's revenue in the third quarter fell by approximately 14.6% (volume -19%) in the desktop processor segment, but still grew by 2.3% (volume +8%) in the laptop processor segment.

And yet, as I previously wrote, Intel quietly continues to hold more than 80% of the x86 market in the consumer segment in terms of revenue - $7.9 billion for the 3rd quarter, against a measly $1.5 billion for AMD, although the latter grew by 46 compared to the 2nd quarter % and by 42% year on year.

Please explain how Intel cores are more efficient than AMD cores!

It will be even funnier if in the end Apple absorbs Intel, and Qualcomm is its (Intel) shadow, AMD...and everything will return to "normality" in the presence of pseudo-competition, only with 2 new players who accumulate all the developments in x86/arm (cisc/risc).

One is stronger, the other is much weaker, but necessary for the strong, main player, for minimal attention from the antimonopoly authorities...

As soon as technological processes reached a flat curve and the smartphone market began to stagnate, the whales of this market then turned their attention to another market, where they could potentially, with their accumulated strength and weight, take away customers from the withering companies Intel and AMD. Nothing personal, just want to eat them...

Somewhere on the side Mediatek is still dangling in the hole, but apparently it will suffer the same fate as VIA, as when it was on the x86 market...

As I wrote earlier, Intel and AMD have the best hardware/software translators of x86 code into RISC core code, because All of their processors are internal RISC processors.

But with the gradual loss of significance of the old x86 code base, the erosion of influence on the market, the significance of both, at some fatal moment for both companies, will collapse on the market and the market will immediately forget about them, as has happened more than once in the history of mankind.

This has very serious consequences both for older developers for these platforms and for students currently studying for the developer profession (mostly we are talking about the system level, of course) - it is important in advance, from youth, not to get on board the Titanic...this will determine their entire future career and success.

Isn't it funny that the situation between Arm SoC manufacturers - Apple vs Qualcomm exactly repeats the "confrontation" Intel vs AMD over the past 8 years? As always, Intel has higher single-core performance, but weaker multi-core performance - for the obvious reason - worse technical process. But Apple and Qualcomm are formally in the same league - both have direct access to "3nm" technical processes. It turns out that the Qualcomm team has not yet been able to achieve normal performance at the level of one productive core, but it takes them in quantity, and with the same technical process this is only possible by high consumption of the entire SoC. No miracles.

Obviously. that AMD is technologically inferior to Intel, since every time it loses in single-core performance, despite clearly more advanced technical processes. Gaining in multi-threaded work due to the difference in consumption of the entire SoC, i.e. It is clear to any IT expert that the Intel processor development team makes much more efficient cores, but simply cannot add many of them, because... how their technology processor support team is worse than TSMC/Samsung's. But having its own factories allows Intel to firmly maintain its overwhelming superiority in the x86 processor market in a 5:1 ratio compared to AMD. The latter is simply not able to increase its market share due to lack of access to large and cheap production - TSMC/Samsung cannot provide this to them.

There is also a funny correlation between the Apple vs Qualcomm opposition. Apple is always one step ahead in chartering major volumes to TSMC/Samsung compared to Qualcomm.

It turns out that the winner will be the one who can provide himself not only with the performance of his cores approximately equal to that of his competitor, but also will receive the desired level of access to mass and cheap production of chips, winning in this regard over his competitor. It is not enough to have the technologically best kernels (product) - it is important to have access to huge mass production of them using the most advanced technical processes before a competitor and in GREATER volumes, for LESS (per piece) money...

Intel already lives only on government subsidies; it has been bankrupt for a long time. And AMD is moving towards this, because... it clearly has less chance of entering the Arm SoC market for mass PC/laptop models, because her team is also apparently weaker at the moment. in this regard, like Intel, compared to the burgeoning teams at Apple and Qualcomm.

Quote from: Mr Majestyk on November 02, 2023, 01:29:59Apple [...] for a laptop the battery life is massively better and why I would buy an M3 based laptop anyday over even AMD's best Phoenix offerings. Intel's power usage is a joke.

As a general statement, this is false. There are AMD or Intel notebooks with almost 16h WLAN battery life. Ryzen 7840U is capabable of reaching some 23h battery life with 99.9Wh.

So this is compared to a base M3 with 4 perf and 4 efficiency cores. The Snapdragon has 12 cores that are all alike (performance).

Wait until the M3 Max (12 performance cores) benchmarks come out for it to be a fair comparison, assuming it'll only show the 12 performance cores in the benchmark.

The M3 pro with 6+6 core configuration may even beat the SD and if so that goes to show you who really is the multicore winner.

Seeing overall geekbench results, it's pretty much a 13900H (slightly faster single core slightly slower multicore) minus the heat and battery drain.

And yes people care about not burning their hands while using a laptop. Still curious about how those cut-down m3 pro perform. Feels like the pro line drew the short end of the stick this time

Quote from: Mr Majestyk on November 02, 2023, 01:29:59

Quote from: Bob John on November 02, 2023, 01:14:35All of these scores are abysmal compared to competing high end x86 CPUs.

LOL sure at what power usage. I don't Apple but for a laptop the battery life is massively better and why I would buy an M3 based laptop anyday over even AMD's best Phoenix offerings. Intel's power usage is a joke. Why do you think Apple got rid of the Intel rubbish in the first place.

Maybe Meteor Lake will be a huge improvement, but they are only talking 30% less power than cRaptor Lake, which will still see it use way more power than the Apple chips.

Lol, who cares about battery life, we aren't in the dark ages anymore you can get a big external battery for cheap and you are very likely to find a place to plug your charger in. You also don't have to use your laptop at full power the entire time you are unplugged.

Quote from: Bob John on November 02, 2023, 01:14:35All of these scores are abysmal compared to competing high end x86 CPUs.

LOL sure at what power usage. I don't Apple but for a laptop the battery life is massively better and why I would buy an M3 based laptop anyday over even AMD's best Phoenix offerings. Intel's power usage is a joke. Why do you think Apple got rid of the Intel rubbish in the first place.

Maybe Meteor Lake will be a huge improvement, but they are only talking 30% less power than cRaptor Lake, which will still see it use way more power than the Apple chips.