Intel Core i9-10885H appears slower than the i7-10875H, by almost 20 percent

Pete Flint

Posts: 28   +7
Staff
In a nutshell: The Intel Core i9-10885H has been found running 20 percent slower than expected in Dell Precision laptops, capping its clock speed at just 2.7GHz. Most likely down to the limited cooling in this entry-level Dell device, this is a reminder to consumers and professionals to investigate performance in laptops and PC builds as a whole before investing in them.

Intel's vPro line of CPUs are business-class chips, typically aimed at IT professionals designed with onboard hardware security firmware. These features defend against attacks below the OS and allow companies to repair and manage their machines remotely.

In a recent laptop benchmark comparison ran by NotebookCheck, the vPro-enabled Core i9-10885H has seen results nearly 20 percent slower than the Core i7-10875H, its consumer-oriented equivalent. With the same core count and cache, and boasting 200MHz higher boost speeds, the i9-10885H wins on paper. That said, benchmarks on the laptop CPU as found in the Dell Precision 3551 showed roughly 20 percent less performance in multi-threaded tasks, and closer to a 10 percent deficit in benchmarks like 7-Zip, Blender, and LibreOffice.

At 100 percent utilization, the i9-10885H stabilized at just 2.7GHz, only 300 MHz above its base clock, so this 20 percent difference can most likely be attributed to Dell's cooling solution in the Precision 3551, causing bottlenecks in the CPU. This is an entry-level workstation laptop, coming in below the 5000 and 7000 series, and priced at around $2,100.

This is another example where performance can vary heavily between machines with the same or similar hardware and emphasis shouldn't just be placed on individual components. If you want the best performance for your consumer or workstation laptop, the build as a whole needs to back up the parts you're paying for.

Permalink to story.

 

seeprime

Posts: 503   +544
If we needed more evidence that Intel's manufacturing tech is still years behind it's own earlier schedules, this is it. They seem to have maximized what they can do on their five year old 14-nm process and can't get their 10-nm process to run optimally, so their faster chips still tend to be 14-nm.

AMD's 20% consumer market share is only going to climb further.
 

Theinsanegamer

Posts: 2,061   +2,617
"At 100 percent utilization, the i9-10885H stabilized at just 2.7GHz, only 300 MHz above its base clock, so this 20 percent difference can most likely be attributed to Dell's cooling solution in the Precision 3551, causing bottlenecks in the CPU"

HARD disagree here. If this were the case, there would likely be 0 performance difference, because the throttling is far below the boost speed of either CPU. If thermals were the limitation then both CPUs woudl throttle to roughly the same point. A delta this large smells to me like a bad factory paste job or faulty heatsink/mounting pressure. vPRO doesnt take that much power to run.

We've seen this at work, most of our 7390 2 in 1s are fine, but occasionally you get one with poor mounting pressure or a totally botched paste job that overheats constantly. We even had one that has the screws for the heatsink totally stripped out.

We also had one that was built fine, but just ran WAY hotter thent he others. We figured that one was a CPU silicon lottery loser that somehow made it to i5 level instead of being binned as an i3.
 

Squid Surprise

Posts: 3,565   +2,462
Um... we need a LOT more information before drawing ANY conclusions. Was this a one-off benchmark with just one 10885 laptop? Were numerous laptops tested?

This could be simply due to the one laptop being a lemon... as a previous poster said, anything can happen in the manufacturing process to mess up a laptop...

Will there be a follow up article with any information? Or is this just an attempt to smear Intel CPUs?

From the ORIGINAL article:
" It's unlikely that we'll get to test many more Core i9-10885H laptops as this particular processor is quite uncommon, but it's something to keep in mind for those who want to squeeze as much performance as possible out of their vPro-enabled business laptops or mobile workstations. "

So this is just a one-off.... let's get some more testing done please!
 

Bulllee

Posts: 67   +43
Dell sales must be well up since working from home became the norm.Be plenty of those machines out there weighed off as tax relief.
 

texasrattler

Posts: 1,015   +452
If we needed more evidence that Intel's manufacturing tech is still years behind it's own earlier schedules, this is it. They seem to have maximized what they can do on their five year old 14-nm process and can't get their 10-nm process to run optimally, so their faster chips still tend to be 14-nm.

AMD's 20% consumer market share is only going to climb further.
This has nothing to do with the article to begin with.
Also Intel has been on 14+++ and it still kept up and lead in various categories against AMDs best. Meanwhile its only now that AMD (4 gens later) that they can finally say they will match or beat Intel. But Intel will also launch new models next year so whatever lead AMD has wont last long. Intel has a massive lead to begin with, I doubt they are worried.
 

jgraham11

Posts: 30   +63
This has nothing to do with the article to begin with.
Also Intel has been on 14+++ and it still kept up and lead in various categories against AMDs best. Meanwhile its only now that AMD (4 gens later) that they can finally say they will match or beat Intel. But Intel will also launch new models next year so whatever lead AMD has wont last long. Intel has a massive lead to begin with, I doubt they are worried.
I doubt that. With Intel using a 14nm process and now only transitioning to 10nm. TSMC will be on 5nm next year... I wouldn't doubt Zen 3 is re-released using the new node. I don't think Intel will be able to compete. Your point about it being 4 gens later for AMD to finally catch up: Some would argue that Intel's performance advantage was at the cost of power, typically drawing double what the rated TDPs to be able to compete on performance. The 9900k's (95WTDP) power use boosted to 140Watts. The 10900k (125W) boosts to almost 200W. Meanwhile the 3800X (105W TDP) may boost to 110W. The 3700x is even more extreme at 65W, boosting to 80W max.

Look at Renoir: the 4800u is in a league of its own destroying Intel completely.

This article doesn't surprise me...
 
  • Like
Reactions: Charles Olson

Toju Mikie

Posts: 130   +132
It could just be that the i9 is not designed to go into slim laptops. The only place they should be in are thicker laptops with good cooling. Macbooks also had the same issue where the i9 version had lower performance than the i7 version. Basically people that bought the Macbook with the i9 were paying more money for less performance.
 

Gomos

Posts: 43   +70
I can say that this is absolutely not true in the Precision 7000 series, because I own a 7750 laptop with the i9-10885H and I have compared my benchmarks with those of other Precision 7750 users, with either the i7 or the i9. Normally the i9 is slightly faster (very slightly, likely not worth the price). In any case, this must be a problem only with the Precision 3000 series.
 
  • Like
Reactions: Stryc

Lew Zealand

Posts: 1,631   +1,688
TechSpot Elite
It could just be that the i9 is not designed to go into slim laptops. The only place they should be in are thicker laptops with good cooling. Macbooks also had the same issue where the i9 version had lower performance than the i7 version. Basically people that bought the Macbook with the i9 were paying more money for less performance.
Yup, I confirmed this with my own testing, so when it came time for Mrs. Zealand to get a new MBPro for work, she got the i7.
 

Mr Majestyk

Posts: 503   +420
What sort of dumb f@$ks would cram that heat pump into a cramped laptop. This is truly hilarious and Dell and Intel should be ashamed of their incompetence as should anybody stupid enough to buy one.
 
  • Like
Reactions: Irata and ruddevil

scavengerspc

Posts: 707   +516
TechSpot Elite
Dell. No surprise to me. In spring 2019 I was looking for a new desktop replacement and one of the machines I tried was the Area 51m. It ran well but after a few minutes of really frantic action, it started choking. It got really hot really fast and only finding a place to cover or pausing would cool it down though not for long. They ended up with a "fix", a bios update, that lowered the 2080 from 200 watts to 170 and jacked the fan speeds dramatically. Max performance was gone and soon nobody was talking about the 51m anymore but people that bought in earlier were $4600 worth of screwed.

 
  • Like
Reactions: Charles Olson

moon982

Posts: 43   +6
If we needed more evidence that Intel's manufacturing tech is still years behind it's own earlier schedules, this is it. They seem to have maximized what they can do on their five year old 14-nm process and can't get their 10-nm process to run optimally, so their faster chips still tend to be 14-nm.

AMD's 20% consumer market share is only going to climb further.

Intel marketing has never been clear. The golden rule i3 is for lower end systems, i5 for mid rage system and i7 for high end systems and i9 for really really high end systems don’t always play out with intel marketing.

There some i3s that run faster than i5s and i5s that run faster than i7s.

Blame Intel marketing and praying on some one getting computer book and learning computers and making sense of tech jargon.

Likewise every generation is faster than the older generation is not always true example some 4 GEN i7 running faster than 8 GEN i7.

Again blame Intel marketing team.

The Intel generation and number means really nothing. You really have look at CPU and compare it to other CPU and look at the benchmarks not going by numbering and generation and that higher the number the better it is.

Again blame Intel marketing.

The i9 quoted here is probably not really i9 but i7 it just intel marketing team is calling it that.


Next time you go to computer store and there is 100 computers for sale get pen and paper write it down than go home and compare it and look at the benchmarks.

It said but the numbers and generations don’t really mean much.

One of my computers have i3 and base on the benchmarks it runs faster than most i3s and lot of i5s it really i5 but intel calls it i3.

The numbers and generations don’t really mean much.

Blame Intel marketing team.
 
"At 100 percent utilization, the i9-10885H stabilized at just 2.7GHz, only 300 MHz above its base clock, so this 20 percent difference can most likely be attributed to Dell's cooling solution in the Precision 3551, causing bottlenecks in the CPU"

HARD disagree here. If this were the case, there would likely be 0 performance difference, because the throttling is far below the boost speed of either CPU. If thermals were the limitation then both CPUs woudl throttle to roughly the same point. A delta this large smells to me like a bad factory paste job or faulty heatsink/mounting pressure. vPRO doesnt take that much power to run.

We've seen this at work, most of our 7390 2 in 1s are fine, but occasionally you get one with poor mounting pressure or a totally botched paste job that overheats constantly. We even had one that has the screws for the heatsink totally stripped out.

We also had one that was built fine, but just ran WAY hotter thent he others. We figured that one was a CPU silicon lottery loser that somehow made it to i5 level instead of being binned as an i3.
Dell confirmed that the Precision 3551 uses a 35 W CPU design while the Latitude 7550 is a 45 W design which explains why our Core i9 Precision is perform slower than a Core i7 Latitude.)