News Intel Meteor Lake Core Ultra 7 1002H Leaked in Geekbench

[Admin]

A purported Meteor Lake Core Ultra 7 1002H device showed up in Geekbench.

Intel Meteor Lake Core Ultra 7 1002H Leaked in Geekbench : Read more

 

[bit_user]

Hey, that's not bad. 4.5% faster 1T @ 96% of the clock speed. If we divide out the clockspeed difference, we get a 8.9% IPC improvement. It's not huge, but it's roughly on par with the improvement Apple claims for its A17.

The multithreaded situation is a little more worrisome (95.7%), given that it's up by 2/2 cores/threads. However, these are probably their "ultra efficiency" cores, in the I/O tile, and therefore don't add a whole lot. Perhaps our clue that something's not right with the Intel 4 node should be the difference in base clocks (1.7 GHz vs. 3.4).

Anyway, wasn't Raptor Refresh rumored to hold the top-spec laptop CPUs? If true, that would be a warning that Intel 4 has some clock scaling problems. I'm finding this reminiscent of the Ice Lake vs. Whiskey Lake situation.

 

[JamesJones44]

Hey, that's not bad. 4.5% faster 1T @ 96% of the clock speed. If we divide out the clockspeed difference, we get a 8.9% IPC improvement. It's not huge, but it's roughly on par with the improvement Apple claims for its A17.

The multithreaded situation is a little more worrisome (95.7%), given that it's up by 2/2 cores/threads. However, these are probably their "ultra efficiency" cores, in the I/O tile, and therefore don't add a whole lot. Perhaps our clue that something's not right with the Intel 4 node should be the difference in base clocks (1.7 GHz vs. 3.4).

Anyway, wasn't Raptor Refresh rumored to hold the top-spec laptop CPUs? If true, that would be a warning that Intel 4 has some clock scaling problems. I'm finding this reminiscent of the Ice Lake vs. Whiskey Lake situation.

I had similar thoughts. I was also curious to what if anything the thermal envelope is playing. Are these both running at 45W/115W or is one set with a lower thermal limit? Even for that matter are they using the same chassis/cooling system. Maybe the 1002H is throttling more in multi thread tests due having a lesser cooling system.

The low 1.7 GHz clock rate is odd though, seem like there could be more to this story. One thing that is odd is the article says:

Other specs revealed during the software’s probing of the system were that the Ultra 7 1002H ran at a base / boost of 3.4 / 5.0 GHz

Which doesn't match what is in the table. However, the GB screenshot shows the max frequency to be 4.98 GHz so I wonder if the table or the article has typos (or both).

 

[kwohlt]

Anyway, wasn't Raptor Refresh rumored to hold the top-spec laptop CPUs?

RPL should only be used in the HX laptop chips, which have always just been desktop chips repackaged for laptop. Can't have a MTL based HX chip if there's no MTL desktop dies.

 

[bit_user]

RPL should only be used in the HX laptop chips, which have always just been desktop chips repackaged for laptop. Can't have a MTL based HX chip if there's no MTL desktop dies.

Tiger Lake had its own HX chip. To your point, that had all the signs of being a desktop CPU that simply never reached the desktop. Maybe it just got further along, before Tiger Lake-S got cancelled. The alternative of putting Rocket Lake in any kind of laptop also would've been a bit daft, so maybe the 8-core TGL was kept alive simply for the HX, and not due to any desktop ambitions.