Intel’s Tiger Lake 11th Gen Core i7-1185G7 Review and Deep Dive: Baskin’ for the Exotic

253 Comments

View All Comments

• JayNor - Thursday, September 17, 2020 - link

  I noted from Intel's Thunderbolt 3 documents that the ports are bidirectional and can, for example, support pcie send while receiving display port on the same cable.
  
  Is it possible, for example, to use an external GPU card with one cable to display output on your laptop's display?

• Spunjji - Friday, September 18, 2020 - link

  I understand that you can, but it drops performance significantly.

• Spunjji - Thursday, September 17, 2020 - link

  Oof, those GPU benchmarks are painful. It's become clear that their predictions were all based on LPDDR5

• PeachNCream - Thursday, September 17, 2020 - link

  Yes the GPU results are somewhat disappointing, but there is only so much you can do when sharing bandwidth to RAM with the CPU cores and everything else. Of course, there is also significant latency to contend with when you don't have GDDR5/6 available to the graphics processor.

• Spunjji - Friday, September 18, 2020 - link

  Absolutely - it's become pretty clear that's why AMD decided to go with 8 CUs on Renoir. I wasn't expecting anything huge from Xe on TGL, but Intel were pushing it as a big win and really it's just not - at least, not in this form. A ~20% bandwidth boost might well translate into big gains on later devices.

• JayNor - Saturday, September 19, 2020 - link

  false, they used lpddr4x for the benchmarks.

• Spunjji - Saturday, September 19, 2020 - link

  I think perhaps you misread what I said. Intel were previously showing numbers suggesting a huge leap in performance for their Xe iGPU which isn't borne out by the testing done here. I'm suggesting that it's because Intel were quoting numbers based on an LPDDR5 implementation. They could have just been lying, though. You seem to be suggesting the latter?

• undervolted_dc - Thursday, September 17, 2020 - link

  Ok but why there are no power consumption of the ryzen ? I see a 28w tiger lake which have a peak of 50+ and an average of 35 and another "28w" tiger lake which instaed have a peak of 50+ and an average of 38w..,
  is the 4800u in thermal throttling because the tiger lake is better cooled ? this "reference" and the lack of real power usage comparision is "strange" to me... we started from the intel benchmarks.. and now we hare here ... where will be in 1 more months when the tiger lake will be in stores with also it's prices.. will be still comparable to the 4800u? ( and I'm also sure that asus will not cover air intake for the tiger lake one to be able to sell them against the 4800u given the higher prices they have to ask .. )
  and where we will be then when zen3/cezanne will be revealed ? you will see the zen3 ipc and freq gain in 1 month.. ( maybe even before tiger lake approach stores ) .. I see no Baskin here..

• undervolted_dc - Friday, September 18, 2020 - link

  Also ram is unfair for this comparison:
  intel LPDDR4X-4266
  vs amd DDR4-3200
  
  but again.. at the end of the day the only thing that matter is the price/performance/power-usage balance for laptop, and here I see only performance comparison..(in unknown thermal/power condition for tests) and with no words about price..
  
  we see here a coming-soon quad core intel with a power usage higher than 1 year old 8 core amd which is also faster in full speed tests ( not in gpu , but it's the old navi chip )
  
  yes, single core bench the intel win.. but their high freq single core are 20% higher freq than AMD ones, are 20% faster, and eat 50% more power , are ~1 year newer , and probably will costs 50% more.. so who is the real winner?

• undervolted_dc - Friday, September 18, 2020 - link

  a mere 4500u with lpddr4x show an average +17% improvements in benchmarks..
  
  https://optocrypto.com/amd-ryzen-4000-adding-suppo...