• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD demonstrates Ryzen 9 5900X prototype with 3D V-Cache stack chiplet design

Cyber-Mav said:
Can they destroy intel though? Amd is loosing market share by the day and intel now holds the performance and value crown. Its getting dire
Click to expand...

The issue AMD have, and will always have, is supply. Intel can just produce more fabs (there are several currently in construction) that are obviously, exclusively for Intel's use. Now that Intel 10nm (Intel '7') is fixed and working well, Intel can produce millions of wafers.

Meanwhile, AMD have to beg/pay increased costs to TSMC, competing with the likes of Apple.

Even if AMD were to defeat Intel in ST and MT with Zen4, Intel will still outsell them by a ridiclous margin, simply because Intel has supply.

I'd love to see AMD get their own fabs back, though fear global foundries are too far behind to be practical. AMD of course can't afford to buy TSMC.
 
Dave2150 said:
The issue AMD have, and will always have, is supply. Intel can just produce more fabs (there are several currently in construction) that are obviously, exclusively for Intel's use. Now that Intel 10nm (Intel '7') is fixed and working well, Intel can produce millions of wafers.

Meanwhile, AMD have to beg/pay increased costs to TSMC, competing with the likes of Apple.

Even if AMD were to defeat Intel in ST and MT with Zen4, Intel will still outsell them by a ridiclous margin, simply because Intel has supply.

I'd love to see AMD get their own fabs back, though fear global foundries are too far behind to be practical. AMD of course can't afford to buy TSMC.
Click to expand...

Producing more fabs doesn't cost anything.
 
dont know why anyone would want Intel or AMD to be destroyed

both with a 50/50 market share with one pushing it to 40/60 every now and then is the best thing for users it keeps things fresh and pushes things
 
Game said:
dont know why anyone would want Intel or AMD to be destroyed

both with a 50/50 market share with one pushing it to 40/60 every now and then is the best thing for users it keeps things fresh and pushes things
Click to expand...
Exactly!

And 50/50 on average with 40/60 followed by 60/40 the ideal. That's why I sadly sort of agree that if anything AMD need a longer run at the top even if shorter-term this means higher prices. The ideal long-term for the consumers (assuming x86 is still the only game in town and no third player emerges) is for AMD to have a good run; get to 60%; then for Intel to come back have a few years good run; get back to 60% and so on.

While not happy about it, I would be willing to pay a bit more to stop Intel killing AMD's growth with predatory pricing before AMD hit the high 40% marketshare.

What I am most concerned about is the chance to make a mistake:

Intel due to their crazy marketshare and margins were over and over again able to afford tons (P4, Larrabee, Itanium, Atom contra revenue, telecoms (low end they are okay but with the billions the poured in they shouldn't have been happy to supplant just 3Com but really should be above Cisco and Juniper), McAfee, 4G/5G modems, etc.

AMD's two major blunders (overpaying for ATI and Bulldozer) nearly killed them.

In the P4 area, Intel were able to keep the P3 / Pentium M projects going eventually giving us Dual Core.

For a longer-term balanced market, AMD must have deep enough pockets to be able to afford a P4/Bulldozer but have a CAT type core as the backup plan.

As for ADL vs Zen3 Veneer efficiency. It is in very contrived circumstances - basically when ADL can boost well for ST performance of certain/most games. However, it does demonstrate that AMD cannot afford to not only make the Renoir APU follow-up a "budget" part. They have to offer a full-fat cache heavy monolith design too.
 
KompuKare said:
Exactly!

And 50/50 on average with 40/60 followed by 60/40 the ideal. That's why I sadly sort of agree that if anything AMD need a longer run at the top even if shorter-term this means higher prices. The ideal long-term for the consumers (assuming x86 is still the only game in town and no third player emerges) is for AMD to have a good run; get to 60%; then for Intel to come back have a few years good run; get back to 60% and so on.

While not happy about it, I would be willing to pay a bit more to stop Intel killing AMD's growth with predatory pricing before AMD hit the high 40% marketshare.

What I am most concerned about is the chance to make a mistake:

Intel due to their crazy marketshare and margins were over and over again able to afford tons (P4, Larrabee, Itanium, Atom contra revenue, telecoms (low end they are okay but with the billions the poured in they shouldn't have been happy to supplant just 3Com but really should be above Cisco and Juniper), McAfee, 4G/5G modems, etc.

AMD's two major blunders (overpaying for ATI and Bulldozer) nearly killed them.

In the P4 area, Intel were able to keep the P3 / Pentium M projects going eventually giving us Dual Core.

For a longer-term balanced market, AMD must have deep enough pockets to be able to afford a P4/Bulldozer but have a CAT type core as the backup plan.

As for ADL vs Zen3 Veneer efficiency. It is in very contrived circumstances - basically when ADL can boost well for ST performance of certain/most games. However, it does demonstrate that AMD cannot afford to not only make the Renoir APU follow-up a "budget" part. They have to offer a full-fat cache heavy monolith design too.
Click to expand...

I think AMD actually made a mistake not keeping Bulldozer on the backburner with a small CPU design team keeping new versions ticking over. It was actually an OK CPU for cheap laptops compared to the Intel Atom dross at the time, and a testament to that was with Intel's Apollo Lake and Gemini Lake where AMD could keep recooking the same 28nm Bulldozer chip at boosted frequences to match. If they kept it going on older process nodes than Zen as a backup they would have something to fall back on if Zen went wrong. Similar to how the Pentium M was a mobile chip adopted as the basis for the Core 2 Duo.
 
Monkeynut said:
I think AMD actually made a mistake not keeping Bulldozer on the backburner with a small CPU design team keeping new versions ticking over. It was actually an OK CPU for cheap laptops compared to the Intel Atom dross at the time, and a testament to that was with Intel's Apollo Lake and Gemini Lake where AMD could keep recooking the same 28nm Bulldozer chip at boosted frequences to match. If they kept it going on older process nodes than Zen as a backup they would have something to fall back on if Zen went wrong. Similar to how the Pentium M was a mobile chip adopted as the basis for the Core 2 Duo.
Click to expand...
+1 to all that except I though it was the Cat team which should have been kept on!

If AMD are going to follow Intel in copying ARM's big.LITTLE ideas, then it will probably be something like Zen2 cores + Zen 4/5 cores. Or basically Zen "Current" as the big, and Zen "last gen" as the little cores. A Zen2/3 hybrid might not make sense yet since the two cores aren't that noticeably different in size.
 
KompuKare said:
+1 to all that except I though it was the Cat team which should have been kept on!

If AMD are going to follow Intel in copying ARM's big.LITTLE ideas, then it will probably be something like Zen2 cores + Zen 4/5 cores. Or basically Zen "Current" as the big, and Zen "last gen" as the little cores. A Zen2/3 hybrid might not make sense yet since the two cores aren't that noticeably different in size.
Click to expand...

Fair point about Bobcat which I'd forgotten about. AMD are really taking a big risk putting all their eggs in Zen though. We saw this with the Athlon 64 which reached its eventual conclusion with the Phenom II (though I think they should have stretched it further rather than going with Bulldozer and they probably wouldn't have reached disaster). AMD needs a backup CPU in case it all goes to hell and Intel makes a few big leaps, which they have managed in the past because of their size.
 
You must log in or register to reply here.
Back
Top Bottom