Intel 11th Generation Core Tiger Lake-H Performance Review: Fast and Power Hungry
by Brett Howse & Andrei Frumusanu on May 17, 2021 9:00 AM EST- Posted in
- CPUs
- Intel
- 10nm
- Willow Cove
- SuperFin
- 11th Gen
- Tiger Lake-H
SPEC CPU - Multi-Threaded Performance
Moving onto multi-threaded SPEC CPU 2017 results, these are the same workloads as on the single-threaded test (we purposefully avoid Speed variants of the workloads in ST tests). The key to performance here is not only microarchitecture or core count, but the overall power efficiency of the system and the levels of performance we can fit into the thermal envelope of the device we’re testing.
It’s to be noted that among the four chips I put into the graph, the i9-11980HK is the only one at a 45W TDP, while the AMD competition lands in at 35W, and the i7-1185G7 comes at a lower 28W. The test takes several hours of runtime (6 hours for this TGL-H SKU) and is under constant full load, so lower duration boost mechanisms don’t come into play here.
Generally as expected, the 8-core TGL-H chip leaves the 4-core TGL-U sibling in the dust, in many cases showcasing well over double the performance. The i9-11980HK also fares extremely well against the AMD competition in workloads which are more DRAM or cache heavy, however falls behind in other workloads which are more core-local and execution throughput bound. Generally that’d be a fair even battle argument between the designs, if it weren’t for the fact that the AMD systems are running at 23% lower TDPs.
In the floating-point multi-threaded suite, we again see a similar competitive scenario where the TGL-H system battles against the best Cezanne and Renoir chips.
What’s rather odd here in the results is 503.bwaves_r and 549.fotonik_r which perform far below the numbers which we were able to measure on the TGL-U system. I think what’s happening here is that we’re hitting DRAM memory-level parallelism limits, with the smaller TGL-U system and its 8x16b LPDDR4 channel memory configuration allowing for more parallel transactions as the 2x64b DDR4 channels on the TGL-H system.
In terms of the overall performance, the 45W 11980HK actually ends up losing to AMD’s Ryzen 5980HS even with 10W more TDP headroom, at least in the integer suite.
We also had initially run the suite in 65W mode, the results here aren’t very good at all, especially when comparing it to the 45W results. For +40-44% TDP, the i9-11980HK in Intel’s reference laptop only performs +9.4% better. It’s likely here that this is due to the aforementioned heavy thermal throttling the system has to fall to, with long periods of time at 35W state, which pulls down the performance well below the expected figures. I have to be explicit here that these 65W results are not representative of the full real 65W performance capabilities of the 11980HK – just that of this particular thermal solution within this Intel reference design.
Facebook
Twitter
RSS
229 Comments
View All Comments
Yojimbo - Monday, May 17, 2021 - link
Should read "And frankly, I can't see a good reason for many consumers to be looking at the situation differently such that they would be concerned about the power draw of the laptop when plugged in."I'm not saying that plugged in power usage is useless to consider, just that it seems to me much less important, as far as power-usage is concerned (even for a desktop replacement) than battery-powered power usage and performance. Maybe others feel differently but I don't understand why. It's not like any laptop is a real power hog unlike some desktop systems can be. We're talking about, what, plus-or-minus 20 watts here? 30 watts? 30 watts plugged in means nothing to me. Does it mean a lot to most others, and if so why?
Bik - Monday, May 17, 2021 - link
It's the capability of the laptop to disperse heat. More watt = more heat. The heat is the issue (loud fan, cpu throttle).Yojimbo - Monday, May 17, 2021 - link
But that varies widely from laptop to laptop. And it's not just a function of the heat output, it's a function of the cooling system, which is related to both cost and weight. So you don't really get helpful information for noise or throttling just by looking at plugged-in power usage.vegemeister - Tuesday, May 18, 2021 - link
More power = more noise for the same cost and weight, or more cost and weight for the same noise.repoman27 - Monday, May 17, 2021 - link
Is there a reason why TGL-U is referred to (somewhat confusingly) in the article as just TGL? I know Intel (also somewhat confusingly) uses the TGL-U 4+2 LP die for three separate platforms (UP3, UP4, and H35), but they're all still considered TGL-U, aren't they? Whereas Tiger Lake is the codename for the whole range of processor families including UP3, UP4, H35, and H.Maybe TGL 4+2 and TGL 8+1 would be more succinct?
Andrei Frumusanu - Monday, May 17, 2021 - link
Fair, I'll change the terminology.bernstein - Monday, May 17, 2021 - link
@Andrei Frumusanu :how is it that the amd 5800x (and others) spec2017fp_r results differ by as much as 7 points while the spec2017int_r values are basically on point? (there seem to be minor differences for the 49/4800U in spec2017int_r values too. plus large diff for the i9-10900K).
comparing to : https://www.anandtech.com/show/16252/mac-mini-appl...
Andrei Frumusanu - Monday, May 17, 2021 - link
In that article we were using only the C/C++ sub-benchmarks due to not having a functioning Fortran compiler on the M1 at the time. So it's apples-and-oranges in terms of the scores between the articles. The integer suite only has 1 Fortran workload, the FP suite has much more.Since, I've rerun the M1 ST scores on a vanilla LLVM and Gfortran toolchain to get all workloads, and anyhow all articles except for that initial M1 piece have the full subset of workloads. The M1 MT scores are missing from this piece as I never ran that (brainfart) and no longer have an M1 system at hand.
bernstein - Monday, May 17, 2021 - link
thanks for clearing that up. great articles btw!also thx for anticipating (and answering) my next question!
Ppietra - Monday, May 17, 2021 - link
As far as I know the results in the Mac mini review aren’t the full SPEC 2017, because some tests require a Fortran compiler that didn’t exist for the M1.