-
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.
Core 9000 series
- Thread starter orbitalwalsh
- Start date
It does certainly follow their model but I think some including myself were surprised because of the fact that their plans have had to change in the last 18 months. Previously, every generation had the same top-tier CPU (4c/8t) with the same TDP and power requirements. This time around they've got a 6 core mainstream flagship on their Z370 series and, supposedly, a new 8 core mainstream flagship that will also work on Z370. That was never a given.
I'd say it's both forward and backward as Z170 also supported both Skylake and Kaby Lake, but then to all intents and purposes they're the same CPU
.However, I take your point that Z270 saw no upgrade path.
@DragonQ To be fair, the jump to 6 and 8 cores is bigger than any of their previous gen jumps and they did use a lot of the reserved pins on the socket for power delivery, so at least this time it's not just +/- 1 pin 
I reckon they could probably use 1151v2 for the next gen of CPUs but that's wishful thinking.
I reckon they could probably use 1151v2 for the next gen of CPUs but that's wishful thinking.
Soldato
Yes, I know, I just got my nodes mixed up.
Intel intended to shrink the chipset down to 14nm for Z390 and bring all functionality in-house, but 14nm capacity was already overloaded. So Intel stuck with 22nm.
https://www.overclock3d.net/news/cp...tedly_scrapped_and_replaced_by_z370_rebrand/1
And therein was my point: if Z390 is ultimately just using the same hardware as Z370 and how rushed the 9000 series release has been, there's no conceivable way Intel could make changes to lock 9000 to Z390 at a hardware level. They could've BIOS locked the 9700 and 9900 to Z390 because "power phases, yo", but we're already seeing 8-core confirmation from vendors.
So my original point was Intel haven't offered backward compatibility because of any example set by AMD, they've done so because they literally can't stop it.
I mean, you're just using a rumor mill article, which most likely is the typical fud before any launch.
Also I'm pretty sure that it's the socket and implicitly power delivery that have a lot more impact on what CPUs can be used than the chipset, unless they improved the DMI link (which they probably will with Icelake). They seemingly made a few changes to power delivery with 1151v2, they used a lot of the reserved pins the socket had for it.
Also I'm pretty sure that it's the socket and implicitly power delivery that have a lot more impact on what CPUs can be used than the chipset, unless they improved the DMI link (which they probably will with Icelake). They seemingly made a few changes to power delivery with 1151v2, they used a lot of the reserved pins the socket had for it.
Soldato
Oh I know it's a rumour mill article, as with everything else we discuss ever when it comes to unreleased products, but which of these two Intel thought processes are more plausible?
"Oh hell, AMD have committed to AM4 until 2020 and the prospect of a proper upgrade path is boosting their sales even more. I think we better do the same".
or
"Oh hell, we can't build Z390 on 14nm. Is there any way we can lock the 9000s out when repurposing Z370? No? Damn. Next time maybe..."
"Oh hell, AMD have committed to AM4 until 2020 and the prospect of a proper upgrade path is boosting their sales even more. I think we better do the same".
or
"Oh hell, we can't build Z390 on 14nm. Is there any way we can lock the 9000s out when repurposing Z370? No? Damn. Next time maybe..."
I mean, Z390 will most likely have additional features integrated, like WiFi ac MAC, Bluetooth, USB 3.1 & SDXC support, it being a Z370 rebadge makes no sense, just because they use the same process doesn't mean they're going to be the same.
The main reason they're probably not doing Z390 on 14nm is because capacity in their 14nm fabs is limited since most of their product stack is already using that process. 10nm geting delay after delay is the reason for that; thee was news a while ago that they were cutting H310 production (14nm) for the same reason.
The main reason they're probably not doing Z390 on 14nm is because capacity in their 14nm fabs is limited since most of their product stack is already using that process. 10nm geting delay after delay is the reason for that; thee was news a while ago that they were cutting H310 production (14nm) for the same reason.
Soldato
That is exactly what I said
And all the other features due for Z390 are done with 3rd party controllers rather than in-house stuff.
That is exactly what I said
And all the other features due for Z390 are done with 3rd party controllers rather than in-house stuff.
It's going to be fairly easy to see if it's a new chipset or not when the board specs are out, Z390 has integrated WiFi ac MAC & Bluetooth so we should see specific chips being used. Similar to how motherboard makers have to use Intel I211-AT ethernet controllers for AMD boards because the AMD chipset doesn't integrate the ethernet MAC like Z370 does, where they just use a PHY like the I219-V.
Associate
- Joined
- 31 Aug 2017
- Posts
- 2,209
So another day another vuln, at this rate intel will need a 10 Ghz 8 core to compete once all the mitigations are in place to patch up there sive like cores, welcome Foreshadow.
They're all the same class of vulnerabilities and the fixes don't really impact consumer workloads. This L1TF vulnerability mainly impacts VMs.
They're supposed to have hardware mitigation with Cascade Lake so it's going to be interesting to see how's that going to affect performance.
They're supposed to have hardware mitigation with Cascade Lake so it's going to be interesting to see how's that going to affect performance.
Pretty much, with more to come too.
Just saying that Intel has been sleeping since around Sandy Bridge's launch. Those incremental IPC gains and frequency gains caused by finer lithography are laughable.
Soldato
Sleeping or milking? Huge difference.
For all of AMDs gains, Intel are still the gaming king. Who is to say they can't keep that up as and when they need to. God knows what they have been working on and sticking on the back shelves while they milked us.
Caporegime
Same arguments were made when Ryzen was just a rumour.
2 years on sales between AMD and Intel are about 50/50 and Intel are still rehashing the same CPU's, again.
Intel have nothing.
Soldato
Gaming king true but security shiv also. And if all motherboard manufacturers apply the Intel patches on the exploits, there will be a 25% perf drop.
Already there are benchmarks testing each one of the tests and the perf goes down up to 8% on average per exploit patch.
Soldato
Got a source for this 25% gaming performance loss?
Soldato
Unfortunately there isn't a single benchmark place that combines all for the desktop, as many manufacturers haven't pushed out all the patches yet.
Each place compares only one patch at the time on average is 2-3% per patch, some tested the V4 and drop on it's own 8%.
On the Xeon side though where there are more solid work on benchmarks (google search), showing a 13-21% drop depending the task, as of May 2018. Before the last round of updates.
Soldato