HP ProBook 450 15.6 inch G9 benchmarks (Core i5-1235U)

Jul 27, 2020
16,817
10,764
106
Managed to spend some quality time with this laptop at the office. It was going to someone else but I asked the IT guy to let me have it for a few hours.

First, let's get the physical attributes out of the way. The build is pretty sturdy. It looks and feels like a hunk of metal. Thanks to being fairly heavy, it would definitely work as a murder weapon with a blow to the head. The silver metallic (probably some alloy) casing doubles as a heatsink and radiates the heat of the internal components pretty well. I think I heard the fans ramp up for a few seconds once or twice at most, despite benchmarking on it for hours. Honestly, that was surprising. The kind of experience I would have expected from an AMD laptop, not an Intel one. So kudos to HP on that.

The display wasn't anything out of the ordinary. Just a run of the mill 45% NTSC, as far as I could tell. Possibly IPS (don't remember angling the lid too much to get a better view). The keyboard isn't backlit sadly, something I would have expected at this price point ($842 for 16GB DDR4-3200 RAM, Geforce MX 570 and 1 TB Crucial P3 SSD). Turns out that our regular shop providing laptops to our office gave us a good deal. Other retailers are selling this laptop for the same price with 8GB RAM, 512GB SSD and Geforce MX 550. The keyboard itself is nothing to write home about. Gets the job done. Really wished it was better. Touchpad tapping was fine but the buttons were too hard for my liking.

Didn't have time to update Windows to the latest version.

1672236359375.png

CPU-Z screenshots of installed memory:

cpuz slot1.pngcpuz slot2.png

RAM speed was dynamically switching 1T and 2T profiles:

1T.png2T.png

MaxxMem² score: https://forums.anandtech.com/threads/post-your-maxxmem²-score.2608515/post-40919186

Had to be careful inserting a USB flash drive and wired mouse into the USB ports. They were not easy to plug into. I think this is a problem plaguing most metallic laptop builds, as I faced the same issue with my old ASUS N550JV.

Plugged in and on battery power plan was set to High Performance.

The dGPU:

gpu.png

Windows was freshly installed by IT and he forgot to install the Geforce driver so downloaded it from Nvidia's site and installed it. The MX 570 GPU showed up in Device Manager and was also used by Unigine Superposition on the first run and after enabling Max DC performance (see below for more on that):

unigine.pngunigine max DC.png

The dGPU has 2GB of its own VRAM but it can use up to 6GB from the system RAM for a total of 8GB. Couldn't do anymore testing with Unigine Superposition on subsequent reboots post Max DC performance enabled as Unigine stopped seeing the dGPU and would only show the iGPU.

Out of box Geekbench 5 score: https://browser.geekbench.com/v5/cpu/19537637

I went into the BIOS looking to disable the E-cores but HP doesn't provide that option. With only two P-cores in this CPU, maybe Intel and/or HP don't want support call headaches from people insisting that their CPU had 12 cores yesterday and suddenly 8 of those have vanished! I did find something called Max DC performance (something to do with plugged in turbo boost?) so I enabled it and checked GB5 again: https://browser.geekbench.com/v5/cpu/19538591

Max DC perf setting impact on ST: +1.96%
Max DC perf setting impact on MT: -4.69%

I decided to leave this setting on as the ST score is comparable to the Core i5-12400 and it's nice to know that I'm getting desktop class ST performance. The only other thing I could change now was to turn off HT so here's the new GB5 score: https://browser.geekbench.com/v5/cpu/19539457

Comparison with stock
Disable HT impact on ST: +5.63%
Disable HT impact on MT: -18.07% (!)

Comparison with HT on and Max DC perf
Disable HT impact on ST: +3.6%
Disable HT impact on MT: -14.04%

No matter how you look at it, disabling HT on the two P-cores has a significant impact on MT performance. I guess in most cases, HT should be left on, at least for this CPU.

Next up is Rapydmark on High setting: https://forums.anandtech.com/threads/rapydmark-cpu-benchmark.2599240/post-40919168

CPU-Z benchmarks show it being comparable to i7-8700K desktop CPU.

Max DC HT enabled
1672239690501.png

Max DC HT disabled

1672241246599.png

Disabling HT affects both ST and MT here!
 
Jul 27, 2020
16,817
10,764
106
Max DC HT enabled on battery

1672241571936.png

18.229% decrease in ST and 8.5% decrease in MT on battery.

Let's disable HT too.

1672241746545.png

Hmm..ST increased very slightly while MT decreased by only 1.118%. Should be interesting how this impacts the battery life. Too bad I can't find out.
 
Jul 27, 2020
16,817
10,764
106
Stock GB5: https://browser.geekbench.com/v5/cpu/19537637

Max DC Performance enabled: https://browser.geekbench.com/v5/cpu/19538591

dGPU scores, all with Max DC:


OOPS! Looks like I couldn't get the Vulkan score for MX 570:
 
Jul 27, 2020
16,817
10,764
106
BMARK

AC Power - Battery
bmark dc max high perf HT on.pngbmark dc max high perf HT on battery.png

Speedometer 2.0

With HT - HT disabled - HT disabled on battery
speedometer HT on dc max high perf.pngspeedometer HT off dc max high perf.pngspeedometer HT off dc max high perf battery.png

Both tests done in MS Edge Chrome.
 

Starjack

Member
Apr 10, 2016
25
0
66
I went into the BIOS looking to disable the E-cores but HP doesn't provide that option. With only two P-cores in this CPU, maybe Intel and/or HP don't want support call headaches from people insisting that their CPU had 12 cores yesterday and suddenly 8 of those have vanished! I did find something called Max DC performance (something to do with plugged in turbo boost?) so I enabled it and checked GB5 again: https://browser.geekbench.com/v5/cpu/19538591

Neither does my Core i3-1215U powered Lenovo laptop has the option to disable E-cores in the BIOS. Is probably would be the same with most 12th Gen laptops. But as seen with most of your benchmarks, it doesn't make not that much of difference in performance anyway. I even watch a video on YouTube recently in which the user play games with E-cores enabled and disabled on a desktop 12th Gen CPU and still, the games run at similar frame rates.
 

LightningZ71

Golden Member
Mar 10, 2017
1,631
1,901
136
The 1235U is such an odd duck. At several of the review sites for laptops, the aggregate performance of the 1235U is typically within a couple percent of the I5-1135G7/1155G7. The E-cores run very slowly, topping out at 3.3Ghz. Directly comparing to the 1155G7, which is within 100Mhz of P core boost speed, you can see the MT advantage of having 8 e cores instead of 2 P cores with HT. The MT scores are typically notably higher for AlderLake. Now, to control for as much as we can, compare the 1255U to the 1165G7. Same L3, same boost clocks, same iGPU EU count. You see the architectural improvement of the P cores come through in the St tests (along with the larger L2), but you really get to see the MT improvements in several tests with the E cores (which boost a bit higher), though, some of that can also be due to the DDR5 ram that's sometimes configured in the systems.

The bigger issue here is that there just aren't enough P cores to begin with. In all of the discussions about keeping HT or optimizing for ST throughput, it was pointed out that there really needs to be 4-6 P cores for good system performance. However, you'd think that adding 2 more P cores would make a big difference in the MT scores, especially if HT threads were really that impactful... but it doesn't! Compare the 1235U to the 1240P. Same L3, same P core boost, same E core boost, but 2 more P cores on the 1240P. You'd think that, with 2 more P cores, and two more HT threads which are SO impactful, that the 1240P would just run away with things... The 1235U manages to outscore it often and often straight up wins comparison benchmarks. Same TDP settings for both to boot. There are other limitations at play here.
 
  • Like
Reactions: igor_kavinski