Rahh
[H]ard|Gawd
- Joined
- Jan 14, 2005
- Messages
- 1,607
I'm cheering AMD on. Really would like to see them a power house int he chip and gpu sectors!
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
AMD Thinks Intel's 10nm Stumble will Open Doors
- Thread starter FrgMstr
- Start date
2019 will likely been a great year for AMD CPU's. I'm excited to see what Ryzen 3000 can do. However, Intel has more cash and they have Jim Keller now. Anyone that thinks AMD wouldn't have done the same thing Intel did without competition, you're lying to yourself. As for GPU's I hope Intel does make come out with a great one. I would welcome another GPU to choose from.
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,941
Since AMD is vulnerable as well, I don't see this part as an advantage to AMD. Although both have promised fixes in the next gen (which now seems like 2019 for both).
Rockenrooster
Gawd
- Joined
- Apr 11, 2017
- Messages
- 955
I believe MELTDOWN is where the big performance hit is, and AMD is not susceptible to MELTDOWN, only Spectre.
I am not sure if either side will be too concentrated on hardware fix. Reason being they fix it you will see someone sit there and hammer on the cpu to find other vulnerabilities. I am not saying they wont fix it I am just saying they are likely not motivated to fix it in hardware anytime soon. You will likely see them fix this sort of stuff by software patches.
drescherjm
[H]F Junkie
- Joined
- Nov 19, 2008
- Messages
- 14,941
Remember that both Intel and AMD promised a hardware fix in the next gen. Whether or not these fixes remove the vulnerability or just make it more difficult is very unclear. I would lean towards the latter.
I think they will probably limit the performance hit with software and also we all know someone then will start hammering the CPUs to find another bug for 2-3 years and then we may get similar news again lol. Always seems to happen.
Yea, I meant Meltdown not Spectre, although at this point, who's really counting?
Oh how short a memory people have. People said that about Intel before when AMD's Athlon 64 was king. You have to remember that Intel has dominated AMD for most of the last two and a half decades or so. While things are certainly getting interesting, Intel still has many advantages over AMD. Money, R&D budget, facilities, etc. all come into play. The only reason AMD had success with the Athlon / Athlon 64 was due to a series of corporate acquisitions and the hiring of key personnel laid off from the then recent DEC layoffs after Compaq bought them. AMD has only had this shot is because Intel has been complacent for so long due to a total lack of competition from anyone.
One thing is for sure, this could be very good for us and things are only going to continue to get interesting from here on out.
I think the difference this time is that Intel apparently banked on their 10nm process getting fixed in 2018 and since that didn't materialize they're going to be caught with their pants down for something like nearly 2yr or so. They won't be in as bad of position as AMD was with Bulldozer vs Sandybridge but there is real potential for this to be much worse than the P4 vs A64 days for them. Rumor mill is saying its just going to be minor Coffeelake refreshes, higher TDP's for greatly diminishing clock speed returns, and maybe some core count bumps on 14nm++(+) and that is it for that whole time period on consumer desktop. Server space won't be looking all that much better either.
On top of that process scaling is something that is becoming so much more problematic and the gains more incremental that Intel might've just permanently blown their typically huge multi year process advantage they enjoyed for so long. With no real major process advantage to count on anymore they're going to have to compete on only design...and price. Which is something they probably haven't been forced to do in a long long time (back when AMD beat Intel to a copper process I think?) and certainly not in a "this is the new normal" fashion.
I don't think AMD can topple Intel in ~2yr here but AMD has a real chance to come closer to evening the score than they've had since A64 vs P4 and making that sort of gain stick too. I think one of their biggest limitations was always production (IIRC they could only supply something like a third of global x86 CPU demand at best...which meant Intel was always guaranteed the other 2/3's almost no matter what) and now they can go to GF or TSMC to make their chips they can supply more of the x86 market than they ever could.