10 ms·

This seems absolutely inevitable. If IFS is serious it needs to be fabbing SoC's with Arm CPUs and to be as competitive as possible with TSMC. Nothing forces

This seems absolutely inevitable. If IFS is serious it needs to be fabbing SoC's with Arm CPUs and to be as competitive as possible with TSMC. Nothing forces IFS to make Arm CPU's that compete with Intel's desktop and server products. Of course, the two companies have a long history of collaboration including Intel having an Arm architecture license via StrongARM. There is an interesting anecdote from Robin Saxby of Arm about when Steve Jobs phoned him and tried to persuade him to block the sale of StrongARM to Intel but Saxby convinced Jobs otherwise.

hashtag-til 4y ago
This is the definitive comment in this thread ^.
pavlov 4y ago
Intel sold off Xscale (née StrongARM) just before the touchscreen smartphone boom started. Seems like they would have been well positioned to take the lead in this market. Probably Apple would have used an Xscale SOC for the iPhone if Intel had taken it seriously. Strategic blunder of the century so far?
- baybal2 4y ago
  Non-touchscreen smartphone boom been going for years prior. The problem was that embedded OS-es like Symbian, A200, PalmOS were running circles around Windows Mobile. Most of them used Samsung SoCs which were few times slower than StrongARM. StrongARM was seen exclusively in the context of it being used to "lift" extremely bloated WinMo smartphones, and was otherwise thought as an overkill for other uses.
- baq 4y ago
  The biggest mistake of Paul Ottelini for sure. The world would look very, very different if he said yes.
  - lr1970 4y ago
    Before becoming Intel's CEO Paul Otellini served as executive VP of sales and marketing. This uber-blunder is a great example what happens when you let a "bean-counter" run a tech company. EDIT: corrected factual error. Paul was never a CFO but much of his career was in sales and marketing. Sales do count as "bean-counters" because their main metric are counting PaL.
    - klelatti 4y ago
      That's not correct. Otellini did not serve as CFO at any time and did not have a finance background. Please check your facts before making a comment like this. Edit: I see you've changed your comment without acknowledgment. Still wrong 'sales and marketing' isn't 'bean counting' and Otellini had extensive 'tech' experience.
      - lr1970 4y ago
        Thanks for pointing out a factual error. I corrected it and added explicit EDIT paragraph.
    - voidfunc 4y ago
      Have their been any successful major companies run by former CFOs? My experience here is that it is often a death sentence via a long road or stagnation.
    - bushbaba 4y ago
      Disagree. The Sales leaders are generally also looking for continued growth 3 years out. Sales know there’s sprints but you are also running a marathon where continued growth YoY is expected. They should have seen the smartphone potential and started chipping away to grow their base long term. Similar to why the best sales leaders only do deals where there’s renewal possibilities in 2-3 years out.
    - svnt 4y ago
      I mean you should probably just own the whole mistake — in no world are bean-counters sales people. https://dictionary.cambridge.org/dictionary/english/bean-counter https://dictionary.cambridge.org/dictionary/english/bean-cou...
- helsinkiandrew 4y ago
  Xscale sale was announced 7 months before the iPhone was announced, but presumably way after work had started and Apple had asked Intel for a processor. A truly poor decision. > There was just one problem: The PC era was about to end. Apple was already working on the iPhone, which would usher in the modern smartphone era. Intel turned down an opportunity to provide the processor for the iPhone, believing that Apple was unlikely to sell enough of them to justify the development costs. https://www.vox.com/2016/4/20/11463818/intel-iphone-mobile-revolution https://www.vox.com/2016/4/20/11463818/intel-iphone-mobile-r...
  - LeifCarrotson 4y ago
    > 7 months before the iPhone was announced, but presumably way after work had started and Apple had asked Intel for a processor You say that as if Intel should have known that the iPhone would be an explosive hit. They probably assumed it would be yet another Blackberry competitor, not really a threat to their desktop hegemony.
    - helsinkiandrew 4y ago
      They didn't have to know that the iPhone would have been an explosive hit, but they should have seen that Smart Phone/PDA like device sales were increasing at 30-50% a year at that time, and there was a definite feeling that this was the future. Even if the iPhone wasn't an explosive hit and the smartphone growth didn't explode, 10% of Blackberries market would have meant millions of processors sold to Apple in the first years and any other Blackberry competitor that was interested,
      - marcosdumay 4y ago
        > there was a definite feeling that this was the future Some people thought that, yeah. But not everybody, and not on the same places Intel sold their chips to. Anyway, Intel sold their foundry exactly because its unitary profit was way too small for them. Since then, the company only got more expensive, so they wouldn't be able to keep it.
      - 4y ago
- DeathArrow 4y ago
  But Intel did respond with x86 SoCs for smartphones. Merrifield, Moorefield and Airmont.
  - unethical_ban 4y ago
    I still have an Asus android tablet with an Intel CPU, circa 2014. It wasn't that great then and it deserves death now. Zero ability to upgrade or get third party OS on it due to those processors not taking off.
  - Sohcahtoa82 4y ago
    And they were plagued with compatibility and performance problems. Almost bought my niece an Android tablet with an x86 CPU. They were priced super low. Turns out there was a reason for that. Apps ran like shit and any games written using the Android NDK just straight-up didn't work.
- anecdotal1 4y ago
  Intel actually designed a CPU for the iPhone. Apple didn't use it because it failed to meet their performance/thermal requirements. Intel was not happy about this.
  - pavlov 4y ago
    Was it an x86 core rather than Xscale?
  - sharedbeans 4y ago
    Sourcing for this? This is a very under-reported story IMO, the main source we have is self-serving comments from the guy who screwed it up saying “we passed because it wasn’t going to be profitable at the projected volumes”. But I don’t see how Intel would have even been capable of being the sort of partner that mobile SoC development requires.
    - anecdotal1 4y ago
      Consider this a rumor until I can remember the source. There was a deep dive article I came across about this a while ago and the only other interesting bits I've never seen anywhere else were tha - Intel was offering a chip based on core/atom and couldn't hit the required performance/TDP - apple wanted a custom chip meeting the required specs, but this is where Intel balked due to estimates of low demand to design and fab it - apple went with ARM and PowerVR GPU was an important part of this story Sometimes I wish I had a searchable full text index of articles I found interesting but didn't see value in bookmarking
      - bsnnkv 4y ago
        > Sometimes I wish I had a searchable full text index of articles I found interesting but didn't see value in bookmarking I hope you won't mind this reply that isn't related to the topic, but this part of your comment stood out to me. The way I get around this is by just saving the bits of articles I find interesting for later full-text search, and on the side I make topic-specific feeds of what I'm saving available via RSS for others to subscribe to.[1] This paradigm shift from saving/bookmarking entire articles to saving highlights has worked really well for me, maybe it could work for you? [1]: https://notado.app/feeds/jado https://notado.app/feeds/jado - I run the service my highlights and feeds are hosted on, but if you're not interested in a subscription service that does this I don't think it's particularly difficult to spin up something yourself over a weekend or two
- omneity 4y ago
  To me this seems to be an attempt to leverage their dominant position at the time to further strengthen it, and eventually failing to strong-arm the industry.
- wmf 4y ago
  Intel doesn't invest in anything besides x86. They let XScale languish while they owned it and even if the iPhone used XScale, Intel still would have found a reason to under-invest in it.
- oblak 4y ago
  It only cost Otellini his cushy job
- maven29 4y ago
  Blackberry phones peaked in 2009, there was plenty of time to reconsider. This is just blatant revisionism, there were plenty of on-ramp opportunities within the product timeline. Apple didn't move off Samsung entirely until the A8, and they still rely on many Samsung hardware IPs within the SoC to this day. It was not just a fab arrangement from day 1. https://twitter.com/calebccff/status/1472517091970494465 https://twitter.com/calebccff/status/1472517091970494465 Apple relied on many of the same suppliers as the rest of the industry for a very long time, except the ones they acquired and closed-off. Purely on a "rising tide lifts all boats basis", Intel would still have had access to the dominant player if the ARM-based touchscreen era wiped out the incumbents. However, this is still a very low-margin business, despite requiring more engineering effort than desktop. You could argue that $1500 phones were unthinkable at the time, but as a commodity, chipset manufacturers see less than 5-10% percent of that, even on flagship phones.
- samstave 4y ago
  Heh, I was at Intel in the late 90s when they were internally worried as heck of Transmeta (not in relation to LGBTQ or the company meta :-) https://www.zdnet.com/article/arm-shares-tumble-as-transmeta-readies-crusoe/ https://www.zdnet.com/article/arm-shares-tumble-as-transmeta... in 1997 Intel was terrified of Transmeta, and it lit a fire under their ass to get the Celeron out more quickly... Me and a colleague worked a F-ton to test out not only the Celeron, AGP, SIMD, OpenGL, Unreal Engine, and all the latest games... But Intel paid the game devs a Million Marketing bucks if they would optimize their games (and general software) to the SIMD instructions, to make the game feel more subjectively responsive on the Celeron as opposed to the AMD, and the seemingly incoming Transmeta. so I think that Transmeta was the big blunder, but in reverse ; ; Intel should have been all over gaming, and not been afraid of some plucky startup... But recall all the senior engineers at intel from <1995 had slide rules and pocket protectors.... they didnt see the entertainment wave of computing early enough and made far too many mistakes and tried to catch up.... personally, I think Linux saved Intel from bankruptcy.
  - rjsw 4y ago
    Transmeta was attractive to embedded HW developers because it had a PCI host controller built in, you didn't need a separate northbridge chip, the speed was less important.
- rjsw 4y ago
  Xscale didn't have a GPU built into the SoC.
- TradingPlaces 4y ago
  It’s worse than that. Apple had just switched PPC to Intel, and Jobs approached Intel in 2005 to make the iPhone SoC. From Paul Otellini exit interview in 2013 >"We ended up not winning it or passing on it, depending on how you want to view it. And the world would have been a lot different if we'd done it… The thing you have to remember is that this was before the iPhone was introduced and no one knew what the iPhone would do... At the end of the day, there was a chip that they were interested in that they wanted to pay a certain price for and not a nickel more and that price was below our forecasted cost. I couldn't see it. It wasn't one of these things you can make up on volume. And in hindsight, the forecasted cost was wrong and the volume was 100x what anyone thought.” The title of the article is "Paul Otellini's Intel: Can the Company That Built the Future Survive It?” and that is still a good question a decade later. https://archive.is/B0Bbs https://archive.is/B0Bbs
  - ksec 4y ago
    Well that is Intel’s view. Inside Apple their engineers refuse to use anything Intel and threaten Jobs to quit.
baybal2 4y ago
> The collaboration will focus on mobile SoC designs first, but allow for potential design expansion into automotive, Internet of Things (IoT), data center, aerospace and government applications. Take a note on the wording. They are definitely making a mobile SoC for somebody, with Qualcomm being the likeliest client. But then they instantly dilute it with other buzzwords, and try to break the line of thinking how it will compete with Intel's own products.
- klelatti 4y ago
  'Data center' doesn't necessarily mean Xeon competitors.
okdood64 4y ago
From what I understand Arm for servers is not really a thing right now. What sorts of industry shifts would it take for this adoption? What would be the challenges?
- qwytw 4y ago
  Why? AWS has Gravitron and Ampere seems to be growing fast both seem to be very competitive in certain uses cases.
- franga2000 4y ago
  It's very much a thing! AWS and GCP and Oracle have had it for years, probably others as well. Just a few days ago, Hetzner launched Ampere based VPS.
  - okdood64 4y ago
    What are the use cases?
    - franga2000 4y ago
      They're more energy efficient, which means they're cheaper to run. That's about it
systemvoltage 4y ago
IFS needs to branch off of Intel's umbrella as an independent Fab competing with TSMC, vertical integration for Intel has failed to compete in last 5 years.
TylerE 4y ago
I recently went from a massive Intel gaming rig (10000 series i9, 3080, and the 1200w powersupply to match) to a the basic $1999 Mac Studio (albeit joined with the several 2TB external SSDs from the old machine). Night and day... and the Mac is actually faster for programming type stuff and the audio stuff I mess around with - on top of being a much more pleasant environment for both. Even for gaming it's a lot more competent than you'd expect, especially for games that have been ported to Silicon. WoW runs at over 100fps (NB: 1440p, 165hz gaming monitor) with settings turned up pretty high. Factorio runs awesome, and just got a silicon port that bumped performance another 20% or so. Obviously it's not so good at AAA games, but I have a PS5 for that, anyway. A lot more of my Steam library was avilable than I expected.. all the Paradox grand strategy stuff, lots of tycoon/city/base builders... On a day to day level it's just so much more plesant... effectively silent (not recording studio silent, there's a very very slight noise, but it's well below the noise floor if the household HVAC is running, or you have any sound at all playing. Even maxed out (like a chess engine using all cores) the fans barely come off idle and the machine temp peaks in the high 50s. At idle, it sites in the very low 30s, only 10c above ambient - and things that are on the edge of the SoC, like the RAM, are only 5c above ambient. Then old i0 beast would idle in the 60s and go all the way into the 90s under load, with fans sounding like jet turbines.
- a_carbon_rod 4y ago
  I made a similar transition as you earlier this year from a i7 8700k/GTX 1080 Ti Windows desktop to a Mac Studio and much like you I'm very happy with that decision for a lot of the reasons you already described. One of the more interesting aspects of the switch I hadn't considered though is that it actually (positively) impacted my power bill - I'm paying about $5-$10 less per month since the switch. The difference in form factor is welcome as well - some days I find myself astonished at how much power fits in such a (relatively) small box on my desk.
  - faeriechangling 4y ago
    The Mac Studio was a bit of a sleeper. For AI inferencing at 128gb it is higher performance than any affordable consumer solution with that much memory available, it's quiet, it's small, it's power efficient, it supports 5 displays. It costs a lot less than an equivalent MBP.
ChancyChance 4y ago
The only real Arm competitor to Intel servers is Ampere. There are literally thousands of customers to TSMC that Intel can serve, as long as they don't just focus on whales. Gelsinger needs to realize intel isn't in the pole position and hasn't been for a years, but he keeps acting like it.

hashtag-til 4y ago

This is the definitive comment in this thread ^.

pavlov 4y ago

Intel sold off Xscale (née StrongARM) just before the touchscreen smartphone boom started. Seems like they would have been well positioned to take the lead in this market. Probably Apple would have used an Xscale SOC for the iPhone if Intel had taken it seriously. Strategic blunder of the century so far?

baybal2 4y ago
Non-touchscreen smartphone boom been going for years prior. The problem was that embedded OS-es like Symbian, A200, PalmOS were running circles around Windows Mobile. Most of them used Samsung SoCs which were few times slower than StrongARM. StrongARM was seen exclusively in the context of it being used to "lift" extremely bloated WinMo smartphones, and was otherwise thought as an overkill for other uses.
baq 4y ago
The biggest mistake of Paul Ottelini for sure. The world would look very, very different if he said yes.
- lr1970 4y ago
  Before becoming Intel's CEO Paul Otellini served as executive VP of sales and marketing. This uber-blunder is a great example what happens when you let a "bean-counter" run a tech company. EDIT: corrected factual error. Paul was never a CFO but much of his career was in sales and marketing. Sales do count as "bean-counters" because their main metric are counting PaL.
  - klelatti 4y ago
    That's not correct. Otellini did not serve as CFO at any time and did not have a finance background. Please check your facts before making a comment like this. Edit: I see you've changed your comment without acknowledgment. Still wrong 'sales and marketing' isn't 'bean counting' and Otellini had extensive 'tech' experience.
    - lr1970 4y ago
      Thanks for pointing out a factual error. I corrected it and added explicit EDIT paragraph.
  - voidfunc 4y ago
    Have their been any successful major companies run by former CFOs? My experience here is that it is often a death sentence via a long road or stagnation.
  - bushbaba 4y ago
    Disagree. The Sales leaders are generally also looking for continued growth 3 years out. Sales know there’s sprints but you are also running a marathon where continued growth YoY is expected. They should have seen the smartphone potential and started chipping away to grow their base long term. Similar to why the best sales leaders only do deals where there’s renewal possibilities in 2-3 years out.
  - svnt 4y ago
    I mean you should probably just own the whole mistake — in no world are bean-counters sales people. https://dictionary.cambridge.org/dictionary/english/bean-counter https://dictionary.cambridge.org/dictionary/english/bean-cou...
helsinkiandrew 4y ago
Xscale sale was announced 7 months before the iPhone was announced, but presumably way after work had started and Apple had asked Intel for a processor. A truly poor decision. > There was just one problem: The PC era was about to end. Apple was already working on the iPhone, which would usher in the modern smartphone era. Intel turned down an opportunity to provide the processor for the iPhone, believing that Apple was unlikely to sell enough of them to justify the development costs. https://www.vox.com/2016/4/20/11463818/intel-iphone-mobile-revolution https://www.vox.com/2016/4/20/11463818/intel-iphone-mobile-r...
- LeifCarrotson 4y ago
  > 7 months before the iPhone was announced, but presumably way after work had started and Apple had asked Intel for a processor You say that as if Intel should have known that the iPhone would be an explosive hit. They probably assumed it would be yet another Blackberry competitor, not really a threat to their desktop hegemony.
  - helsinkiandrew 4y ago
    They didn't have to know that the iPhone would have been an explosive hit, but they should have seen that Smart Phone/PDA like device sales were increasing at 30-50% a year at that time, and there was a definite feeling that this was the future. Even if the iPhone wasn't an explosive hit and the smartphone growth didn't explode, 10% of Blackberries market would have meant millions of processors sold to Apple in the first years and any other Blackberry competitor that was interested,
    - marcosdumay 4y ago
      > there was a definite feeling that this was the future Some people thought that, yeah. But not everybody, and not on the same places Intel sold their chips to. Anyway, Intel sold their foundry exactly because its unitary profit was way too small for them. Since then, the company only got more expensive, so they wouldn't be able to keep it.
    - 4y ago
DeathArrow 4y ago
But Intel did respond with x86 SoCs for smartphones. Merrifield, Moorefield and Airmont.
- unethical_ban 4y ago
  I still have an Asus android tablet with an Intel CPU, circa 2014. It wasn't that great then and it deserves death now. Zero ability to upgrade or get third party OS on it due to those processors not taking off.
- Sohcahtoa82 4y ago
  And they were plagued with compatibility and performance problems. Almost bought my niece an Android tablet with an x86 CPU. They were priced super low. Turns out there was a reason for that. Apps ran like shit and any games written using the Android NDK just straight-up didn't work.
anecdotal1 4y ago
Intel actually designed a CPU for the iPhone. Apple didn't use it because it failed to meet their performance/thermal requirements. Intel was not happy about this.
- pavlov 4y ago
  Was it an x86 core rather than Xscale?
- sharedbeans 4y ago
  Sourcing for this? This is a very under-reported story IMO, the main source we have is self-serving comments from the guy who screwed it up saying “we passed because it wasn’t going to be profitable at the projected volumes”. But I don’t see how Intel would have even been capable of being the sort of partner that mobile SoC development requires.
  - anecdotal1 4y ago
    Consider this a rumor until I can remember the source. There was a deep dive article I came across about this a while ago and the only other interesting bits I've never seen anywhere else were tha - Intel was offering a chip based on core/atom and couldn't hit the required performance/TDP - apple wanted a custom chip meeting the required specs, but this is where Intel balked due to estimates of low demand to design and fab it - apple went with ARM and PowerVR GPU was an important part of this story Sometimes I wish I had a searchable full text index of articles I found interesting but didn't see value in bookmarking
    - bsnnkv 4y ago
      > Sometimes I wish I had a searchable full text index of articles I found interesting but didn't see value in bookmarking I hope you won't mind this reply that isn't related to the topic, but this part of your comment stood out to me. The way I get around this is by just saving the bits of articles I find interesting for later full-text search, and on the side I make topic-specific feeds of what I'm saving available via RSS for others to subscribe to.[1] This paradigm shift from saving/bookmarking entire articles to saving highlights has worked really well for me, maybe it could work for you? [1]: https://notado.app/feeds/jado https://notado.app/feeds/jado - I run the service my highlights and feeds are hosted on, but if you're not interested in a subscription service that does this I don't think it's particularly difficult to spin up something yourself over a weekend or two
omneity 4y ago
To me this seems to be an attempt to leverage their dominant position at the time to further strengthen it, and eventually failing to strong-arm the industry.
wmf 4y ago
Intel doesn't invest in anything besides x86. They let XScale languish while they owned it and even if the iPhone used XScale, Intel still would have found a reason to under-invest in it.
oblak 4y ago
It only cost Otellini his cushy job
maven29 4y ago
Blackberry phones peaked in 2009, there was plenty of time to reconsider. This is just blatant revisionism, there were plenty of on-ramp opportunities within the product timeline. Apple didn't move off Samsung entirely until the A8, and they still rely on many Samsung hardware IPs within the SoC to this day. It was not just a fab arrangement from day 1. https://twitter.com/calebccff/status/1472517091970494465 https://twitter.com/calebccff/status/1472517091970494465 Apple relied on many of the same suppliers as the rest of the industry for a very long time, except the ones they acquired and closed-off. Purely on a "rising tide lifts all boats basis", Intel would still have had access to the dominant player if the ARM-based touchscreen era wiped out the incumbents. However, this is still a very low-margin business, despite requiring more engineering effort than desktop. You could argue that $1500 phones were unthinkable at the time, but as a commodity, chipset manufacturers see less than 5-10% percent of that, even on flagship phones.
samstave 4y ago
Heh, I was at Intel in the late 90s when they were internally worried as heck of Transmeta (not in relation to LGBTQ or the company meta :-) https://www.zdnet.com/article/arm-shares-tumble-as-transmeta-readies-crusoe/ https://www.zdnet.com/article/arm-shares-tumble-as-transmeta... in 1997 Intel was terrified of Transmeta, and it lit a fire under their ass to get the Celeron out more quickly... Me and a colleague worked a F-ton to test out not only the Celeron, AGP, SIMD, OpenGL, Unreal Engine, and all the latest games... But Intel paid the game devs a Million Marketing bucks if they would optimize their games (and general software) to the SIMD instructions, to make the game feel more subjectively responsive on the Celeron as opposed to the AMD, and the seemingly incoming Transmeta. so I think that Transmeta was the big blunder, but in reverse ; ; Intel should have been all over gaming, and not been afraid of some plucky startup... But recall all the senior engineers at intel from <1995 had slide rules and pocket protectors.... they didnt see the entertainment wave of computing early enough and made far too many mistakes and tried to catch up.... personally, I think Linux saved Intel from bankruptcy.
- rjsw 4y ago
  Transmeta was attractive to embedded HW developers because it had a PCI host controller built in, you didn't need a separate northbridge chip, the speed was less important.
rjsw 4y ago
Xscale didn't have a GPU built into the SoC.
TradingPlaces 4y ago
It’s worse than that. Apple had just switched PPC to Intel, and Jobs approached Intel in 2005 to make the iPhone SoC. From Paul Otellini exit interview in 2013 >"We ended up not winning it or passing on it, depending on how you want to view it. And the world would have been a lot different if we'd done it… The thing you have to remember is that this was before the iPhone was introduced and no one knew what the iPhone would do... At the end of the day, there was a chip that they were interested in that they wanted to pay a certain price for and not a nickel more and that price was below our forecasted cost. I couldn't see it. It wasn't one of these things you can make up on volume. And in hindsight, the forecasted cost was wrong and the volume was 100x what anyone thought.” The title of the article is "Paul Otellini's Intel: Can the Company That Built the Future Survive It?” and that is still a good question a decade later. https://archive.is/B0Bbs https://archive.is/B0Bbs
- ksec 4y ago
  Well that is Intel’s view. Inside Apple their engineers refuse to use anything Intel and threaten Jobs to quit.

baybal2 4y ago

> The collaboration will focus on mobile SoC designs first, but allow for potential design expansion into automotive, Internet of Things (IoT), data center, aerospace and government applications. Take a note on the wording. They are definitely making a mobile SoC for somebody, with Qualcomm being the likeliest client. But then they instantly dilute it with other buzzwords, and try to break the line of thinking how it will compete with Intel's own products.

klelatti 4y ago
'Data center' doesn't necessarily mean Xeon competitors.

okdood64 4y ago

From what I understand Arm for servers is not really a thing right now. What sorts of industry shifts would it take for this adoption? What would be the challenges?

qwytw 4y ago
Why? AWS has Gravitron and Ampere seems to be growing fast both seem to be very competitive in certain uses cases.
franga2000 4y ago
It's very much a thing! AWS and GCP and Oracle have had it for years, probably others as well. Just a few days ago, Hetzner launched Ampere based VPS.
- okdood64 4y ago
  What are the use cases?
  - franga2000 4y ago
    They're more energy efficient, which means they're cheaper to run. That's about it

systemvoltage 4y ago

IFS needs to branch off of Intel's umbrella as an independent Fab competing with TSMC, vertical integration for Intel has failed to compete in last 5 years.

TylerE 4y ago

I recently went from a massive Intel gaming rig (10000 series i9, 3080, and the 1200w powersupply to match) to a the basic $1999 Mac Studio (albeit joined with the several 2TB external SSDs from the old machine). Night and day... and the Mac is actually faster for programming type stuff and the audio stuff I mess around with - on top of being a much more pleasant environment for both. Even for gaming it's a lot more competent than you'd expect, especially for games that have been ported to Silicon. WoW runs at over 100fps (NB: 1440p, 165hz gaming monitor) with settings turned up pretty high. Factorio runs awesome, and just got a silicon port that bumped performance another 20% or so. Obviously it's not so good at AAA games, but I have a PS5 for that, anyway. A lot more of my Steam library was avilable than I expected.. all the Paradox grand strategy stuff, lots of tycoon/city/base builders... On a day to day level it's just so much more plesant... effectively silent (not recording studio silent, there's a very very slight noise, but it's well below the noise floor if the household HVAC is running, or you have any sound at all playing. Even maxed out (like a chess engine using all cores) the fans barely come off idle and the machine temp peaks in the high 50s. At idle, it sites in the very low 30s, only 10c above ambient - and things that are on the edge of the SoC, like the RAM, are only 5c above ambient. Then old i0 beast would idle in the 60s and go all the way into the 90s under load, with fans sounding like jet turbines.

a_carbon_rod 4y ago
I made a similar transition as you earlier this year from a i7 8700k/GTX 1080 Ti Windows desktop to a Mac Studio and much like you I'm very happy with that decision for a lot of the reasons you already described. One of the more interesting aspects of the switch I hadn't considered though is that it actually (positively) impacted my power bill - I'm paying about $5-$10 less per month since the switch. The difference in form factor is welcome as well - some days I find myself astonished at how much power fits in such a (relatively) small box on my desk.
- faeriechangling 4y ago
  The Mac Studio was a bit of a sleeper. For AI inferencing at 128gb it is higher performance than any affordable consumer solution with that much memory available, it's quiet, it's small, it's power efficient, it supports 5 displays. It costs a lot less than an equivalent MBP.

ChancyChance 4y ago

The only real Arm competitor to Intel servers is Ampere. There are literally thousands of customers to TSMC that Intel can serve, as long as they don't just focus on whales. Gelsinger needs to realize intel isn't in the pole position and hasn't been for a years, but he keeps acting like it.