Jump to content

Some Intel CPU Information

vanished

So I gathered some information from the best non-K CPU in each tier (i3, i5, and i7) for each recent mainstream generation from userbenchmark and tabulated it, and also performed calculations to determine additional things, and then I graphed those things.  The results are below:

 

Raw Data

Spoiler

1.png.107ef0e500713cc3405d90f133e84e02.png

 

Hyper-Threading Data

Spoiler

2.png.c957aea383f476c0a8a6a90f0bf087a8.png

 

In case it's not obvious, this is determined by dividing the multi-core score by the single-core score, factoring in the number of cores and the single-core vs. multi-core clock speeds.  i5s are not shown since they would all read 0% (plus or minus a bit due to errors in the data and testing)

 

IPC Data

Spoiler

3.png.d71f59b24962c7fe6ffa03e30f07c876.png

 

This is determined by comparing the scores with clockspeed factored in so only improvements in score / GHz are revealed.

 

The numbers above the bars show how good that generation is compared to Sandy Bridge, and the numbers on the crescents between the bars shows how much that generation improved compared to the previous.  Note that this would be calculated by dividing the relative performance, not subtracting.  If you get different numbers it's due to rounding.

 

Overall Performance Data

Spoiler

 

4.png.3dda23b3bc04772d4216e0d92c0ea6ce.png

 

 

Single-threaded Performance Data

Spoiler

5.png.f3ad5a6904f30e76bc96d95b52b3e372.png

 

And now as a line chart that uses release date as the x-axis

 

5-2.png.1e799caccf8d51042ec8e77be3d103aa.png

 

 

Multi-threaded Performance Data

Spoiler

6.png.4f9425951e79f0b1438d1b9eccf88816.png

 

Again, as a line chart using release date as x-axis

 

6-2.png.17432f0916dbbfb16e71fb847f1b6497.png

 

 

And to finish, perhaps the most insightful chart of all

Spoiler

Untitled.png.3865cd842e860bf396090058363a3e54.png

 

As closing comments, I have to say I'm not 100% confident in the benchmark scores I've used, since userbenchmark averages all results, even the ones that are clearly stupid (high or low performance outliers), but I am confident in my analysis of them.  If someone has more reliable numbers, I'd be happy to plug them in and let the sheet recalculate everything that results.

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to comment
Share on other sites

Link to post
Share on other sites

How much can ram speed affect these results (especially the jump from DDR3 to DDR4)?

Current LTT F@H Rank: 90    Score: 2,503,680,659    Stats

Yes, I have 9 monitors.

My main PC (Hybrid Windows 10/Arch Linux):

OS: Arch Linux w/ XFCE DE (VFIO-Patched Kernel) as host OS, windows 10 as guest

CPU: Ryzen 9 3900X w/PBO on (6c 12t for host, 6c 12t for guest)

Cooler: Noctua NH-D15

Mobo: Asus X470-F Gaming

RAM: 32GB G-Skill Ripjaws V @ 3200MHz (12GB for host, 20GB for guest)

GPU: Guest: EVGA RTX 3070 FTW3 ULTRA Host: 2x Radeon HD 8470

PSU: EVGA G2 650W

SSDs: Guest: Samsung 850 evo 120 GB, Samsung 860 evo 1TB Host: Samsung 970 evo 500GB NVME

HDD: Guest: WD Caviar Blue 1 TB

Case: Fractal Design Define R5 Black w/ Tempered Glass Side Panel Upgrade

Other: White LED strip to illuminate the interior. Extra fractal intake fan for positive pressure.

 

unRAID server (Plex, Windows 10 VM, NAS, Duplicati, game servers):

OS: unRAID 6.11.2

CPU: Ryzen R7 2700x @ Stock

Cooler: Noctua NH-U9S

Mobo: Asus Prime X470-Pro

RAM: 16GB G-Skill Ripjaws V + 16GB Hyperx Fury Black @ stock

GPU: EVGA GTX 1080 FTW2

PSU: EVGA G3 850W

SSD: Samsung 970 evo NVME 250GB, Samsung 860 evo SATA 1TB 

HDDs: 4x HGST Dekstar NAS 4TB @ 7200RPM (3 data, 1 parity)

Case: Sillverstone GD08B

Other: Added 3x Noctua NF-F12 intake, 2x Noctua NF-A8 exhaust, Inatek 5 port USB 3.0 expansion card with usb 3.0 front panel header

Details: 12GB ram, GTX 1080, USB card passed through to windows 10 VM. VM's OS drive is the SATA SSD. Rest of resources are for Plex, Duplicati, Spaghettidetective, Nextcloud, and game servers.

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, sazrocks said:

How much can ram speed affect these results (especially the jump from DDR3 to DDR4)?

according to when linus tested this.... not much at all

Primary System

  • CPU
    Ryzen R6 5700X
  • Motherboard
    MSI B350M mortar arctic
  • RAM
    32GB Corsair RGB 3600MT/s CAS18
  • GPU
    Zotac RTX 3070 OC
  • Case
    kind of a mess
  • Storage
    WD black NVMe SSD 500GB & 1TB samsung Sata ssd & x 1TB WD blue & x 3TB Seagate
  • PSU
    corsair RM750X white
  • Display(s)
    1440p 21:9 100Hz
Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, sazrocks said:

How much can ram speed affect these results (especially the jump from DDR3 to DDR4)?

Good question.  I don't have exact measurements to give you, but generally speaking, the answer depends on the task.  For some things, it really doesn't matter at all.  For some, it can make quite a noticeable difference.  When it comes to gaming, the rule of thumb is that faster RAM helps only if you are experiencing a CPU bottleneck.

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to comment
Share on other sites

Link to post
Share on other sites

look! Now I have 251 bookmarks. 

Ryzen 5 3600 stock | 2x16GB C13 3200MHz (AFR) | GTX 760 (Sold the VII)| ASUS Prime X570-P | 6TB WD Gold (128MB Cache, 2017)

Samsung 850 EVO 240 GB 

138 is a good number.

 

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Ryan_Vickers said:

Hyper-Threading Data

So when it comes to hyperthreading, Skylake i7 > Kaby Lake i7. Right? 

There is more that meets the eye
I see the soul that is inside

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, hey_yo_ said:

So when it comes to hyperthreading, Skylake i7 > Kaby Lake i7. Right? 

the data suggests so, yes, but this is such an oddity and an outlier, I would want to confirm by examining other benchmarks.  That said, userbemchmark is already inherently an average of many tests, so idk

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to comment
Share on other sites

Link to post
Share on other sites

7 minutes ago, Ryan_Vickers said:

the data suggests so, yes, but this is such an oddity and an outlier, I would want to confirm by examining other benchmarks.  That said, userbemchmark is already inherently an average of many tests, so idk

Could we do like a group project to compile results for different benchmarks and different CPUs?

Judge a product on its own merits AND the company that made it.

How to setup MSI Afterburner OSD | How to make your AMD Radeon GPU more efficient with Radeon Chill | (Probably) Why LMG Merch shipping to the EU is expensive

Oneplus 6 (Early 2023 to present) | HP Envy 15" x360 R7 5700U (Mid 2021 to present) | Steam Deck (Late 2022 to present)

 

Mid 2023 AlTech Desktop Refresh - AMD R7 5800X (Mid 2023), XFX Radeon RX 6700XT MBA (Mid 2021), MSI X370 Gaming Pro Carbon (Early 2018), 32GB DDR4-3200 (16GB x2) (Mid 2022

Noctua NH-D15 (Early 2021), Corsair MP510 1.92TB NVMe SSD (Mid 2020), beQuiet Pure Wings 2 140mm x2 & 120mm x1 (Mid 2023),

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, AluminiumTech said:

Could we do like a group project to compile results for different benchmarks and different CPUs?

the more data the better.  The trick is making sure whatever you pick is

-available for all cpus in the list

-reliable and accurate (aa I said, im still not super happy about even the data I already have)

-represents a large number of tests

-was conducted properly (not running the benchmark while a video renders, making aure the cpu hit and maintained the expected clockspeed throughout the test, etc)

well I guess those last two points are sort of subpoints of #2 but it is important

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to comment
Share on other sites

Link to post
Share on other sites

Hmmm I like whats been done here. This could be very useful info for many things. 

Use this guide to fix text problems in your postGo here and here for all your power supply needs

 

New Build Currently Under Construction! See here!!!! -----> 

 

Spoiler

Deathwatch:[CPU I7 4790K @ 4.5GHz][RAM TEAM VULCAN 16 GB 1600][MB ASRock Z97 Anniversary][GPU XFX Radeon RX 480 8GB][STORAGE 250GB SAMSUNG EVO SSD Samsung 2TB HDD 2TB WD External Drive][COOLER Cooler Master Hyper 212 Evo][PSU Cooler Master 650M][Case Thermaltake Core V31]

Spoiler

Cupid:[CPU Core 2 Duo E8600 3.33GHz][RAM 3 GB DDR2][750GB Samsung 2.5" HDD/HDD Seagate 80GB SATA/Samsung 80GB IDE/WD 325GB IDE][MB Acer M1641][CASE Antec][[PSU Altec 425 Watt][GPU Radeon HD 4890 1GB][TP-Link 54MBps Wireless Card]

Spoiler

Carlile: [CPU 2x Pentium 3 1.4GHz][MB ASUS TR-DLS][RAM 2x 512MB DDR ECC Registered][GPU Nvidia TNT2 Pro][PSU Enermax][HDD 1 IDE 160GB, 4 SCSI 70GB][RAID CARD Dell Perc 3]

Spoiler

Zeonnight [CPU AMD Athlon x2 4400][GPU Sapphire Radeon 4650 1GB][RAM 2GB DDR2]

Spoiler

Server [CPU 2x Xeon L5630][PSU Dell Poweredge 850w][HDD 1 SATA 160GB, 3 SAS 146GB][RAID CARD Dell Perc 6i]

Spoiler

Kero [CPU Pentium 1 133Mhz] [GPU Cirrus Logic LCD 1MB Graphics Controller] [Ram 48MB ][HDD 1.4GB Hitachi IDE]

Spoiler

Mining Rig: [CPU Athlon 64 X2 4400+][GPUS 9 RX 560s, 2 RX 570][HDD 160GB something][RAM 8GBs DDR3][PSUs 1 Thermaltake 700w, 2 Delta 900w 120v Server modded]

RAINBOWS!!!

 

 QUOTE ME SO I CAN SEE YOUR REPLYS!!!!

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, Ryan_Vickers said:

the more data the better.  The trick is making sure whatever you pick is

-available for all cpus in the list

-reliable and accurate (aa I said, im still not super happy about even the data I already have)

-represents a large number of tests

-was conducted properly (not running the benchmark while a video renders, making aure the cpu hit and maintained the expected clockspeed throughout the test, etc)

well I guess those last two points are sort of subpoints of #2 but it is important

Unfortunately windows just adds a bunch of oddities for benchmarking on its own. Random scans and usage spikes with background tasks your can't turn off

Primary System

  • CPU
    Ryzen R6 5700X
  • Motherboard
    MSI B350M mortar arctic
  • RAM
    32GB Corsair RGB 3600MT/s CAS18
  • GPU
    Zotac RTX 3070 OC
  • Case
    kind of a mess
  • Storage
    WD black NVMe SSD 500GB & 1TB samsung Sata ssd & x 1TB WD blue & x 3TB Seagate
  • PSU
    corsair RM750X white
  • Display(s)
    1440p 21:9 100Hz
Link to comment
Share on other sites

Link to post
Share on other sites

9 minutes ago, SquintyG33Rs said:

Unfortunately windows just adds a bunch of oddities for benchmarking on its own. Random scans and usage spikes with background tasks your can't turn off

That's why you get a large scan base, it evens it out almost perfectly once you have enough data. Things like Windows crap is generally such a small load on the cpu it doesn't make much of if any difference to the results

Use this guide to fix text problems in your postGo here and here for all your power supply needs

 

New Build Currently Under Construction! See here!!!! -----> 

 

Spoiler

Deathwatch:[CPU I7 4790K @ 4.5GHz][RAM TEAM VULCAN 16 GB 1600][MB ASRock Z97 Anniversary][GPU XFX Radeon RX 480 8GB][STORAGE 250GB SAMSUNG EVO SSD Samsung 2TB HDD 2TB WD External Drive][COOLER Cooler Master Hyper 212 Evo][PSU Cooler Master 650M][Case Thermaltake Core V31]

Spoiler

Cupid:[CPU Core 2 Duo E8600 3.33GHz][RAM 3 GB DDR2][750GB Samsung 2.5" HDD/HDD Seagate 80GB SATA/Samsung 80GB IDE/WD 325GB IDE][MB Acer M1641][CASE Antec][[PSU Altec 425 Watt][GPU Radeon HD 4890 1GB][TP-Link 54MBps Wireless Card]

Spoiler

Carlile: [CPU 2x Pentium 3 1.4GHz][MB ASUS TR-DLS][RAM 2x 512MB DDR ECC Registered][GPU Nvidia TNT2 Pro][PSU Enermax][HDD 1 IDE 160GB, 4 SCSI 70GB][RAID CARD Dell Perc 3]

Spoiler

Zeonnight [CPU AMD Athlon x2 4400][GPU Sapphire Radeon 4650 1GB][RAM 2GB DDR2]

Spoiler

Server [CPU 2x Xeon L5630][PSU Dell Poweredge 850w][HDD 1 SATA 160GB, 3 SAS 146GB][RAID CARD Dell Perc 6i]

Spoiler

Kero [CPU Pentium 1 133Mhz] [GPU Cirrus Logic LCD 1MB Graphics Controller] [Ram 48MB ][HDD 1.4GB Hitachi IDE]

Spoiler

Mining Rig: [CPU Athlon 64 X2 4400+][GPUS 9 RX 560s, 2 RX 570][HDD 160GB something][RAM 8GBs DDR3][PSUs 1 Thermaltake 700w, 2 Delta 900w 120v Server modded]

RAINBOWS!!!

 

 QUOTE ME SO I CAN SEE YOUR REPLYS!!!!

Link to comment
Share on other sites

Link to post
Share on other sites

8 hours ago, 8uhbbhu8 said:

That's why you get a large scan base, it evens it out almost perfectly once you have enough data. Things like Windows crap is generally such a small load on the cpu it doesn't make much of if any difference to the results

Or rather than getting a bunch of clueless people to run it blindly, you have it run only be people who know what they're doing, on systems they know aren't virus-infested, and they run it several times to confirm the results are consistent, etc.  That would be ideal but there's no site like that with enough data for this table.

 

On that topic, one way I test the validity of the data is through something I call "multi-core scaling".  In theory, in a totally synthetic benchmark, running 4 threads on a quad core should yield virtually exactly 4x the score of running 1 thread on that same quad core (assuming you account for any clock speed differences).  If it's out by more than a few percent, it's a sign to me that something is up.  Here are the values for that, using the quad-threaded test scores:

 

Untitled.png.aac4d8ee3a0ef2fefb99b463dfa57fc7.png

 

Not great as you can tell.  There are a number of chips that deviate by up to 10%.  To me, this is a sign of running the test wrong.  If you look closely, you can see that in every generation, it's the i7 that deviates the most.  One possible way to cause results like this would be to not properly run each thread on its own physical core.  If two threads were allowed to momentarily slip onto a single physical core (due to HT), it would hurt performance, and you would see results like these in the end.  I'm not saying that's what happened, but imo it's a very good theory.

 

Now, to be fair, if the benchmark is written correctly, the user won't have to be aware of any of this, but a user that is aware would be able to take additional steps to ensure it runs correctly, regardless of how the benchmark was written.

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, Ryan_Vickers said:

Or rather than getting a bunch of clueless people to run it blindly, you have it run only be people who know what they're doing, on systems they know aren't virus-infested, and they run it several times to confirm the results are consistent, etc.  That would be ideal but there's no site like that with enough data for this table.

 

On that topic, one way I test the validity of the data is through something I call "multi-core scaling".  In theory, in a totally synthetic benchmark, running 4 threads on a quad core should yield virtually exactly 4x the score of running 1 thread on that same quad core (assuming you account for any clock speed differences).  If it's out by more than a few percent, it's a sign to me that something is up.  Here are the values for that, using the quad-threaded test scores:

 

Untitled.png.aac4d8ee3a0ef2fefb99b463dfa57fc7.png

 

Not great as you can tell.  There are a number of chips that deviate by up to 10%.  To me, this is a sign of running the test wrong.  If you look closely, you can see that in every generation, it's the i7 that deviates the most.  One possible way to cause results like this would be to not properly run each thread on its own physical core.  If two threads were allowed to momentarily slip onto a single physical core (due to HT), it would hurt performance, and you would see results like these in the end.  I'm not saying that's what happened, but imo it's a very good theory.

 

Now, to be fair, if the benchmark is written correctly, the user won't have to be aware of any of this, but a user that is aware would be able to take additional steps to ensure it runs correctly, regardless of how the benchmark was written.

hmm yes that would make a lot of sense. It is a good theory to go on. One thing we do know from LTT tests is that so called "Virtualization" of lower end chips with higher end ones is practical if done properly and all with the same method. 

Use this guide to fix text problems in your postGo here and here for all your power supply needs

 

New Build Currently Under Construction! See here!!!! -----> 

 

Spoiler

Deathwatch:[CPU I7 4790K @ 4.5GHz][RAM TEAM VULCAN 16 GB 1600][MB ASRock Z97 Anniversary][GPU XFX Radeon RX 480 8GB][STORAGE 250GB SAMSUNG EVO SSD Samsung 2TB HDD 2TB WD External Drive][COOLER Cooler Master Hyper 212 Evo][PSU Cooler Master 650M][Case Thermaltake Core V31]

Spoiler

Cupid:[CPU Core 2 Duo E8600 3.33GHz][RAM 3 GB DDR2][750GB Samsung 2.5" HDD/HDD Seagate 80GB SATA/Samsung 80GB IDE/WD 325GB IDE][MB Acer M1641][CASE Antec][[PSU Altec 425 Watt][GPU Radeon HD 4890 1GB][TP-Link 54MBps Wireless Card]

Spoiler

Carlile: [CPU 2x Pentium 3 1.4GHz][MB ASUS TR-DLS][RAM 2x 512MB DDR ECC Registered][GPU Nvidia TNT2 Pro][PSU Enermax][HDD 1 IDE 160GB, 4 SCSI 70GB][RAID CARD Dell Perc 3]

Spoiler

Zeonnight [CPU AMD Athlon x2 4400][GPU Sapphire Radeon 4650 1GB][RAM 2GB DDR2]

Spoiler

Server [CPU 2x Xeon L5630][PSU Dell Poweredge 850w][HDD 1 SATA 160GB, 3 SAS 146GB][RAID CARD Dell Perc 6i]

Spoiler

Kero [CPU Pentium 1 133Mhz] [GPU Cirrus Logic LCD 1MB Graphics Controller] [Ram 48MB ][HDD 1.4GB Hitachi IDE]

Spoiler

Mining Rig: [CPU Athlon 64 X2 4400+][GPUS 9 RX 560s, 2 RX 570][HDD 160GB something][RAM 8GBs DDR3][PSUs 1 Thermaltake 700w, 2 Delta 900w 120v Server modded]

RAINBOWS!!!

 

 QUOTE ME SO I CAN SEE YOUR REPLYS!!!!

Link to comment
Share on other sites

Link to post
Share on other sites

How did you manage to find an IPC improvement from skylake to kaby? 

AMD Ryzen R7 1700 (3.8ghz) w/ NH-D14, EVGA RTX 2080 XC (stock), 4*4GB DDR4 3000MT/s RAM, Gigabyte AB350-Gaming-3 MB, CX750M PSU, 1.5TB SDD + 7TB HDD, Phanteks enthoo pro case

Link to comment
Share on other sites

Link to post
Share on other sites

54 minutes ago, 8uhbbhu8 said:

hmm yes that would make a lot of sense. It is a good theory to go on. One thing we do know from LTT tests is that so called "Virtualization" of lower end chips with higher end ones is practical if done properly and all with the same method. 

Yes, and that makes sense - 1 core from a certain generation should look the same, whether it came from an i3, an i7, etc.

8 minutes ago, Coaxialgamer said:

How did you manage to find an IPC improvement from skylake to kaby? 

I don't invent the data, I just report them :P and the data shows an IPC improvement.  Or at least, it suggests one, since the scores went up by more than the clock speed did, and increased IPC is the only logical explanation.

Solve your own audio issues  |  First Steps with RPi 3  |  Humidity & Condensation  |  Sleep & Hibernation  |  Overclocking RAM  |  Making Backups  |  Displays  |  4K / 8K / 16K / etc.  |  Do I need 80+ Platinum?

If you can read this you're using the wrong theme.  You can change it at the bottom.

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×