News Ryzen 7 5700U Flies Past Ryzen 7 4800U in Geekbench 5

[Admin]

AMD's unreleased Ryzen 7 5700U mobile APU has been sighted in Geekbench 5.

Ryzen 7 5700U Flies Past Ryzen 7 4800U in Geekbench 5 : Read more

 

[MasterMadBones]

So it definitely has higher IPC than Renoir, judging from the single-threaded score, but it's not as good as Vermeer. Maybe this is some sort of hybrid of Zen 2 and Zen 3?

 

[digitalgriffin]

All apu's up to this point have been monolithic. There is a sacrifice if cache because of this. Thats a lot of transistors. We also inow one of zen 3's bug uplift comes from its largr unified cache.

It makes sense. We likely wont see a multichip approach for apu until zen 4 and mcm for navi 3.

 

[Makaveli]

All apu's up to this point have been monolithic. There is a sacrifice if cache because of this. Thats a lot of transistors. We also inow one of zen 3's bug uplift comes from its largr unified cache.

It makes sense. We likely wont see a multichip approach for apu until zen 4 and mcm for navi 3.

I agree the 8mb Cache is what is limiting performance.

 

[ThatMouse]

There's hardly any Ryzen 7 4800U laptops to actually purchase. Are they even making more or do we wait a year? lol!

 

[PapaCrazy]

There's hardly any Ryzen 7 4800U laptops to actually purchase. Are they even making more or do we wait a year? lol!

Yes, it seems strange AMD could dominate the entire laptop market yet don't have much of a presence to speak of. Hopefully there will be more options soon. Asus had an exclusivity deal with AMD that was for 6 months. That should expire soon. But it was only for the 4800hs. Which, TBH, is the chip I'm more interested in. Ideally, AMD will release a 5700HS soon with even greater gains over the 4800 and will actually get that chip into a wider variety laptops.

 

[TCA_ChinChin]

Still disappointed AMD didn't unify their naming scheme with this move to all 5000 series chips. I'm not denying the value and performance of the new chips, regardless of the underlying architecture, but its still something I find annoying.

 

[Upacs]

In the meantime, in our company we are still waiting for delivery of six 4700U laptops, because our supplier cannot get stock. This all seems a bit pointless unless AMD can get more chips out of the factory doors.

 

[Rdslw]

In the meantime, in our company we are still waiting for delivery of six 4700U laptops, because our supplier cannot get stock. This all seems a bit pointless unless AMD can get more chips out of the factory doors.

Depends on an hardware picked. I've seen a lot of Asus & HP laptops while others seems to be lacking a bit. I am surprised if your supplier cannot get a specific one you want, did not offer you a replacement.
I still wait for 5800U as laptop I wanted with 4800u was not released in my 3'rd world country.

 

[shady28]

So it definitely has higher IPC than Renoir, judging from the single-threaded score, but it's not as good as Vermeer. Maybe this is some sort of hybrid of Zen 2 and Zen 3?

Ah, I think it just has higher clocks, not higher IPC. Geekbench 1172 vs 1500+ on Tiger Lake and 1700 on the new Apple M1. So the 5700U is a last gen part with a next gen model number => sneaky marketing. Even the multi-core looks lackluster since TGL gets nearly that with 4 cores. This chip is Zen 2.

Anyway, from what I see AMD needs to focus on delivering a product and stop with the announcements.

 

[MasterMadBones]

Ah, I think it just has higher clocks, not higher IPC.

But that is exactly the point I'm trying to make. Single core clock speed only increased by 2.4%, while performance went up 13.7%. There must be something else at work here, since we should be able to assume that each APU was able to sustain its single core boost clock throughout the test.

Tiger Lake (1185G7) has some weird behaviour in this test, showing 22% higher IPC than even the 5700U and a massive 35% more IPC than the 4800U. That is far more than we should expect on average. In AnandTech's SPEC results, Tiger Lake doesn't do this well at all clock-for-clock. Multi-core scores shouldn't even be looked at, because we don't know what TDP the CPUs were running. Intel was very happy to promote Tiger Lake at 28W, so we should expect that the samples that had a 6000 score were running at that TDP.

 

[shady28]

But that is exactly the point I'm trying to make. Single core clock speed only increased by 2.4%, while performance went up 13.7%. There must be something else at work here, since we should be able to assume that each APU was able to sustain its single core boost clock throughout the test.

Tiger Lake (1185G7) has some weird behaviour in this test, showing 22% higher IPC than even the 5700U and a massive 35% more IPC than the 4800U. That is far more than we should expect on average. In AnandTech's SPEC results, Tiger Lake doesn't do this well at all clock-for-clock. Multi-core scores shouldn't even be looked at, because we don't know what TDP the CPUs were running. Intel was very happy to promote Tiger Lake at 28W, so we should expect that the samples that had a 6000 score were running at that TDP.

GeekBench shows tons of 1165G7 (not the top sku) getting well over 1500 single core. By comparison the 4800U is around 1150-1200 on a good day. Yes there are some outliers that hit 1250ish, but there are some 1165G7s getting ~1600 too and a lot of 4800U chips getting under 1000 single core.

That 5700U is not Zen 3, or at least it had better not be reflective of Zen 3 mobile or AMD is going to lose that space (again).

1165G7 results :

https://browser.geekbench.com/v5/cpu/search?q=1165g7

4800U results :

https://browser.geekbench.com/v5/cpu/search?utf8=✓&q=4800U