This comment about responsiveness/microlag is interesting. In a hardware/engineering perspective, what is the reason for this? Is it because ARM CPUs don't have aggressive turbo clocks?yeah it’s fine.
I didn't say that.
One man's wrong doesn't justify another man's error.
it’s not about the turbo clocks in isolation. More turbo would give you less lag, it’s just about the cores. Everybody does frequency ramping. How much you can afford, and how fast you get there, and how often you can do that thermally — or without killing the battery, is the question.
That sounds more like a Windows and/or Adobe issue to me.
But then wouldn't the newer Photoshop on x86 also benefit from the same improvements (I doubt Adobe would maintain/release/support two significantly different versions of their SW especially on the same OS, except perhaps in very early beta releases)?
Unless the original x86 codebase is now just the backend behind some kind of abstraction layer.
What is going on with these atrocious GB6 results?
People are saying it's due to premature drivers/firmware. An update is apparently supposed to drop on June 17th to fix the issues.
Dude comparing it to the "beta" MTL silicon based Core Ultra laptop. No wonder...
Meaning what? Early firmware?
There were good results end of May, so something has regressed. For instance: https://browser.geekbench.com/v6/cpu/6319364
I think firmware can be ruled out at this point. It's the horrible tile latency on MTL slowing things down. It's their first mass market disaggregated tile architecture with a "useless" NPU coz it doesn't meet Microsoft's TOPS requirement. Intel scammed people good. They just wanted to get this thing out ASAP and be done with it, without caring about end users and the longevity of the laptop. Being a tile, shouldn't have been impossible for them to put in a better NPU, unless they manufactured the NPUs before the MTL silicon became usable, in which case, they should've used the NPU with cheaper laptops instead of pushing it in expensive premium designs.
Would be great if they can sell it at $300 max.
He's also saying that he can't differentiate between Arm compiled apps and x86 emulated, just like Rosetta 2. Huge win for MS here.
This is seriously impressive! He said that while watching the Apple keynote and doing web work, the battery went from 90% to 75% in 2 hours. Very promising.
I don't think QCOM intend on selling SoCs with less than 45 TOPs NPU to be honest. Pretty sure even Purwa will meet that Copilot + PC requirement.
I think Arm and QCOM will settle.Well ARM, this will definitely get you to 50% market share in the PC space. It's the same "old" legal battle, but perfectly timed to rain on their own parade.
ARM torpedoes Windows on ARM: Demands destruction of all PCs with Snapdragon X
It just seems too abstruse: ARM could become relevant for PCs for the first time. However, the company itself is torpedoing the venture.www.heise.de
That's Intel problem. They decided to ship first-generation consumer disaggregated SoCs with all of its problems. Consumer will just buy an product and compare to what they have.
Yeah, with the OLED at max brightness and speakers on, for context to the others here. 14” display, I think 70-75wh battery? That’s an extremely good result.
Yep
Yep
Yep lol.
Yep. Strongly suspect the tile is an issue and migration to P cores is being throttled for “battery life”, and probably even then they get put to E Cores at clocks way too low for interactivity way too often. I suspect Lunar Lake being on-die for the LPE E cores and Skymont improving will ease this and allow better ramping too, but by how much? Tough to say.
