SoylentNews is people
SoylentNews
from the what-about-meltdown-and-spectre? dept.
Intel Announces 9th Gen Core CPUs: Core i9-9900K (8-Core), i7-9700K, & i5-9600K
Among many of Intel's announcements today, a key one for a lot of users will be the launch of Intel's 9th Generation Core desktop processors, offering up to 8-cores on Intel's mainstream consumer platform. These processors are drop-in compatible with current Coffee Lake and Z370 platforms, but are accompanied by a new Z390 chipset and associated motherboards as well. The highlights from this launch is the 8-core Core i9 parts, which include a 5.0 GHz turbo Core i9-9900K, rated at a 95W TDP.
[...] Leading from the top of the stack is the Core i9-9900K, Intel's new flagship mainstream processor. This part is eight full cores with hyperthreading, with a base frequency of 3.6 GHz at 95W TDP, and a turbo up to 5.0 GHz on two cores. Memory support is up to dual channel DDR4-2666. The Core i9-9900K builds upon the Core i7-8086K from the 8th Generation product line by adding two more cores, and increasing that 5.0 GHz turbo from one core to two cores. The all-core turbo is 4.7 GHz, so it will be interesting to see what the power consumption is when the processor is fully loaded. The Core i9 family will have the full 2MB of L3 cache per core.
[...] Also featuring 8-cores is the Core i7-9700K, but without the hyperthreading. This part will have a base frequency of 3.6 GHz as well for a given 95W TDP, but can turbo up to 4.9 GHz only on a single core. The i7-9700K is meant to be the direct upgrade over the Core i7-8700K, and although both chips have the same underlying Coffee Lake microarchitecture, the 9700K has two more cores and slightly better turbo performance, but less L3 cache per core at only 1.5MB per.
Intel also announced refreshed 8 to 18 core high-end desktop CPUs, and a new 28-core Xeon aimed at extreme workstation users.
Related:
Intel Teases 28 Core Chip, AMD Announces Threadripper 2 With Up to 32 Cores
AMD Threadripper 2 Available Starting on August 13
(Score: 4, Interesting) by Anonymous Coward on Tuesday October 09 2018, @11:12AM (2 children)
Lets face it: They don't have a solution so they're ignoring the problem.
(Score: 5, Interesting) by ledow on Tuesday October 09 2018, @12:39PM (1 child)
Not really surprising, is it?
All the Intel performance came from speculative execution - you were effectively following all paths after a decision and executing all the things you would do for both paths, but then just throwing away the "not used" result.
That's been shown to be inherently flawed unless you apply the same security to your speculative executions as if you were really executing them anyway.
That's gonna be a performance hit on anything - hence all the fixes are to basically "break" speculative execution, or remove the possibility, when the chips just go back to ordinary chips with no speculative execution. This is why hyper-threading is no longer recommended or even inside most new chips.
You can make a chip, no doubt, that does the speculative execution properly. But that will require more die-space. Which means a complete redesign AND it's probably just easier at that point to make a chip without SE, but with entire extra cores, for the same price.
People were warning about SE for years because they knew something like this would happen. And it's only really a bodge unless you're literally executing all paths "for real" and the "rollback" / contention etc. of the wrong path doesn't affect the speed of the right path at any point.
Pretty much, the more "fancy tricks" you use to make your stuff go faster, the harder it bites further down the road. And without the fancy tricks and pretending, Intel chips are precisely THIS fast, with THIS many cores, and no fancy compiler/processor tricks take place.
The solution can only be a complete redesign of all chips (which they won't have got close to doing), which will result in a larger/hotter chip to do the same job. Or they remove all the fancy junk, at the same cost, and give you slightly more cores at the same speed.
It'll be 5-10 years before Intel chips progress again from a new design. At least. And even other manufacturer's aren't immune, and all because they played fast-and-loose with security when speculatively executing instructions.
(Score: 2) by RamiK on Wednesday October 10 2018, @07:01AM
Designing a chip without speculative execution that performs even half as good for general purpose loads is a hundred billion dollars problem that baffled Intel and HP with the Itanium and remains unsolved. Disabling hyperthreading in favor of more cores only address a very specific speculative execution vulnerability and only to some extent. The other fixes require recompilation and have a significant performance hit.
Staying OoO, Intel/AMD/ARM are more likely to redesign their microarchitectures following SafeSpec [arxiv.org]. But it's still needs researching and is years away.
Regardless, right now, they really don't have a solution that doesn't hit you with huge performance overheads.
compiling...