• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

CPU TDP definition as of Nov-2021??

Joined
Jan 20, 2019
Messages
1,243 (0.66/day)
Location
London, UK
System Name ❶ Oooh (2024) ❷ Aaaah (2021) ❸ Ahemm (2017)
Processor ❶ 5800X3D ❷ i7-9700K ❸ i7-7700K
Motherboard ❶ X570-F ❷ Z390-E ❸ Z270-E
Cooling ❶ ALFIII 360 ❷ X62 + X72 (GPU mod) ❸ X62
Memory ❶ 32-3600/16 ❷ 32-3200/16 ❸ 16-3200/16
Video Card(s) ❶ 3080 X Trio ❷ 2080TI (AIOmod) ❸ 1080TI
Storage ❶ NVME/SSD/HDD ❷ <SAME ❸ SSD/HDD
Display(s) ❶ 1440/165/IPS ❷ 1440/144/IPS ❸ 1080/144/IPS
Case ❶ BQ Silent 601 ❷ Cors 465X ❸ Frac Mesh C
Audio Device(s) ❶ HyperX C2 ❷ HyperX C2 ❸ Logi G432
Power Supply ❶ HX1200 Plat ❷ RM750X ❸ EVGA 650W G2
Mouse ❶ Logi G Pro ❷ Razer Bas V3 ❸ Logi G502
Keyboard ❶ Logi G915 TKL ❷ Anne P2 ❸ Logi G610
Benchmark Scores I have wrestled bandwidths, Tussled with voltages, Handcuffed Overclocks, Thrown Gigahertz in Jail
Hi,

Just wandering, how does AMD and Intel define "TDP" for CPUs in late 2021?

I understand TDP has nothing to do with actual power consumption. I also understand TDP is in reference to heat output. Why its a necessary measurement, beats me (cooler related?)!! I have a loose take on the subject hence looking for some clarification.

Is it even relevant as some posts I've read suggest it's just a baseline indicator and doesn't really have any significant meaning other than slipping on a cooler with the same/higher TDP rating?
 

Toothless

Tech, Games, and TPU!
Supporter
Joined
Mar 26, 2014
Messages
9,223 (2.52/day)
Location
Washington, USA
System Name Veral
Processor 5950x
Motherboard MSI MEG x570 Ace
Cooling Corsair H150i Elite
Memory 4x16GB G.Skill TridentZ
Video Card(s) Powercolor 7900XTX Red Devil
Storage Crucial P5 Plus 1TB, Samsung 980 1TB, Teamgroup MP34 4TB
Display(s) Acer Nitro XZ342CK Pbmiiphx + 2x AOC 2425W
Case Fractal Design Meshify Lite 2
Audio Device(s) Blue Yeti + SteelSeries Arctis 5 / Samsung HW-T550
Power Supply Corsair HX850
Mouse Corsair Nightsword
Keyboard Corsair K55
VR HMD HP Reverb G2
Software Windows 11 Professional
Benchmark Scores PEBCAK
As you wander around..

Yes, it's per cooler specs. It's a bit different between Intel and AMD, so actually reading reviews helps.
 

freeagent

Moderator
Staff member
Joined
Sep 16, 2018
Messages
7,420 (3.67/day)
Location
Winnipeg, Canada
System Name Altered Beast
Processor AMD R9 5900X/5800X3D
Motherboard Asus Crosshair VIII Dark Hero
Cooling Thermalright Phantom Spirit 120 EVO, 1x T30
Memory 2x8 G.Skill Trident Z Royal 3200C14, 2x8GB G.Skill Trident Z Black and White 3200 C14
Video Card(s) Zotac 4070 Ti Trinity OC
Storage WD SN850X 2TB, SN850 1TB, SN770 1TB - Asus Hyper M.2, 2x SN770 1TB
Display(s) LG 50UP7100
Case Fractal Torrent Compact RGB
Audio Device(s) JBL 2.1 Deep Bass
Power Supply EVGA SuperNova 750w G+, Monster HDP1800
Mouse Logitech G502 Hero
Keyboard Logitech G213
VR HMD Oculus 3
Software Yes
Benchmark Scores Yes
Honestly..

I think its both, sorta. Consumption, and output, they have to be relative to each other minus losses throughout the system. But.. I have a grade 8 edumacation so me and advanced mathematics do not get along.

My 105w 5900X can sit there doing nothing and draw 65w from the wall.. but load it up with something a bit intense and she will put out 240w PPT while pulling 300w+ from the wall, GPU doing nothing. Same for my 65w 5600X.. can do about the same doing nothing, but load it up with something hardcore and she will put out closer to 145w PPT while doing something like 265w at the wall with the GPU doing nothing. These are roundabout numbers, not exact but pretty darn close.

As for coolers.. they need to update their standards for measurement. I have a 360w cooler, and on my old X5690 that could actually put out those numbers, it worked great. But on CPU's like my 3770K, 3600XT, 5600X... it didn't do so great. Exact same story with my 320w Le Grand Macho RT. Now on my 275w FC140, its better, but I don't have my x5690 anymore, so I cant really test it with old school muscle. Dammit. The things we do for money :(
 
Joined
May 8, 2020
Messages
578 (0.41/day)
System Name Mini efficient rig.
Processor R9 3900, @4ghz -0.05v offset. 110W peak.
Motherboard Gigabyte B450M DS3H, bios f41 pcie 4.0 unlocked.
Cooling some server blower @1500rpm
Memory 2x16GB oem Samsung D-Die. 3200MHz
Video Card(s) RX 6600 Pulse w/conductonaut @65C hotspot
Storage 1x 128gb nvme Samsung 950 Pro - 4x 1tb sata Hitachi 2.5" hdds
Display(s) Samsung C24RG50FQI
Case Jonsbo C2 (almost itx sized)
Audio Device(s) integrated Realtek crap
Power Supply Seasonic SSR-750FX
Mouse Logitech G502
Keyboard Redragon K539 brown switches
Software Windows 7 Ultimate SP1 + Windows 10 21H2 LTSC (patched).
Benchmark Scores Cinebench: R15 3050 pts, R20 7000 pts, R23 17800 pts, r2024 1050 pts.
Honestly..

I think its both, sorta. Consumption, and output, they have to be relative to each other minus losses throughout the system. But.. I have a grade 8 edumacation so me and advanced mathematics do not get along.

My 105w 5900X can sit there doing nothing and draw 65w from the wall.. but load it up with something a bit intense and she will put out 240w PPT while pulling 300w+ from the wall, GPU doing nothing. Same for my 65w 5600X.. can do about the same doing nothing, but load it up with something hardcore and she will put out closer to 145w PPT while doing something like 265w at the wall with the GPU doing nothing. These are roundabout numbers, not exact but pretty darn close.

As for coolers.. they need to update their standards for measurement. I have a 360w cooler, and on my old X5690 that could actually put out those numbers, it worked great. But on CPU's like my 3770K, 3600XT, 5600X... it didn't do so great. Exact same story with my 320w Le Grand Macho RT. Now on my 275w FC140, its better, but I don't have my x5690 anymore, so I cant really test it with old school muscle. Dammit. The things we do for money :(
if you disable turbo and just run it at its base clock, then it would sit at about 105w according to the spec. The thing is that both amd and intel have some crazy high voltage for the boost clocks, but that is to ensure that every cpu batch will able to stay stable at such voltage and clock combination. But overall as of 2021 intel is in big trouble: they are making pentium 4 again https://www.techpowerup.com/review/intel-core-i9-12900k-alder-lake-12th-gen/20.html. and the max power consuption is truly insane. Even theold 7900X draws less power, and that is 2066 socket.
 

freeagent

Moderator
Staff member
Joined
Sep 16, 2018
Messages
7,420 (3.67/day)
Location
Winnipeg, Canada
System Name Altered Beast
Processor AMD R9 5900X/5800X3D
Motherboard Asus Crosshair VIII Dark Hero
Cooling Thermalright Phantom Spirit 120 EVO, 1x T30
Memory 2x8 G.Skill Trident Z Royal 3200C14, 2x8GB G.Skill Trident Z Black and White 3200 C14
Video Card(s) Zotac 4070 Ti Trinity OC
Storage WD SN850X 2TB, SN850 1TB, SN770 1TB - Asus Hyper M.2, 2x SN770 1TB
Display(s) LG 50UP7100
Case Fractal Torrent Compact RGB
Audio Device(s) JBL 2.1 Deep Bass
Power Supply EVGA SuperNova 750w G+, Monster HDP1800
Mouse Logitech G502 Hero
Keyboard Logitech G213
VR HMD Oculus 3
Software Yes
Benchmark Scores Yes
Honestly.. I have seen 350w+ system usage using Linpack Xtreme with a serious overclock on my X5690. At least double that when the GPU's had to go to work.. more depending on oc. That thing hated 850w PSU's and polar bears. If they say 165w TDP, what they really mean is 300w and you better have good cooling lol.. because that 280-300w will make you wish you did :toast:
if you disable turbo and just run it at its base clock, then it would sit at about 105w according to the spec.


Interesting.. stock is 3700 on a 5900X right? I can run it at 4600 all loads and my fans can get pretty rowdy.. my cooler has a hard time when this CPU is at its limits, either that or its just the CPU being at its limits lol. Overtemp kicks in at around 105 I think.. it isn't hard to do when you are at the top of the voltage/clock range with these Vermeer's. Or any CPU actually. Interesting enough at the limits of my 5600X instead of 95 being the limit, if you are pulling amps it will kick in at 85c. Kinda neat.. self preservation..
 
Joined
Jun 16, 2013
Messages
1,457 (0.37/day)
Location
Australia
if you disable turbo and just run it at its base clock, then it would sit at about 105w according to the spec. The thing is that both amd and intel have some crazy high voltage for the boost clocks, but that is to ensure that every cpu batch will able to stay stable at such voltage and clock combination. But overall as of 2021 intel is in big trouble: they are making pentium 4 again https://www.techpowerup.com/review/intel-core-i9-12900k-alder-lake-12th-gen/20.html. and the max power consuption is truly insane. Even theold 7900X draws less power, and that is 2066 socket.
Intel maybe in "big trouble" with more electricity needed, but the ipc increase with golden cove cores can go some way in justifying the power consumption. In gaming it makes virtually zero difference unless the game is poorly coded & hammering the cpu to high heaven.
 
Joined
Jan 20, 2019
Messages
1,243 (0.66/day)
Location
London, UK
System Name ❶ Oooh (2024) ❷ Aaaah (2021) ❸ Ahemm (2017)
Processor ❶ 5800X3D ❷ i7-9700K ❸ i7-7700K
Motherboard ❶ X570-F ❷ Z390-E ❸ Z270-E
Cooling ❶ ALFIII 360 ❷ X62 + X72 (GPU mod) ❸ X62
Memory ❶ 32-3600/16 ❷ 32-3200/16 ❸ 16-3200/16
Video Card(s) ❶ 3080 X Trio ❷ 2080TI (AIOmod) ❸ 1080TI
Storage ❶ NVME/SSD/HDD ❷ <SAME ❸ SSD/HDD
Display(s) ❶ 1440/165/IPS ❷ 1440/144/IPS ❸ 1080/144/IPS
Case ❶ BQ Silent 601 ❷ Cors 465X ❸ Frac Mesh C
Audio Device(s) ❶ HyperX C2 ❷ HyperX C2 ❸ Logi G432
Power Supply ❶ HX1200 Plat ❷ RM750X ❸ EVGA 650W G2
Mouse ❶ Logi G Pro ❷ Razer Bas V3 ❸ Logi G502
Keyboard ❶ Logi G915 TKL ❷ Anne P2 ❸ Logi G610
Benchmark Scores I have wrestled bandwidths, Tussled with voltages, Handcuffed Overclocks, Thrown Gigahertz in Jail
if you disable turbo and just run it at its base clock, then it would sit at about 105w according to the spec. The thing is that both amd and intel have some crazy high voltage for the boost clocks, but that is to ensure that every cpu batch will able to stay stable at such voltage and clock combination. But overall as of 2021 intel is in big trouble: they are making pentium 4 again https://www.techpowerup.com/review/intel-core-i9-12900k-alder-lake-12th-gen/20.html. and the max power consuption is truly insane. Even theold 7900X draws less power, and that is 2066 socket.

I somewhat understood the above from intels definition of TDP but dunno for some reason found it hard to digest. Intel informs TDP as:

"Thermal Design Power (TDP) represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload. Refer to Datasheet for thermal solution requirements"

So, base clock/frequency is when all overclocking/boosting functions are disabled and the amount of power dissipated (translated as heat) is what is being referred to as TDP? If correct, I take it I don't need to worry about TDP for gaming with the CPU boost/auto-overclock enabled, right? I always get premium coolers with the likes of 280mm AIOs cooling which I suspect are rated well above the base clock TDP value - correct me if i'm losing my mind lol

EDIT: Funny thing is I've been over this before (probably in TPU forums) but lost a few gazillion brain cells to have forgotten it all.
 
Joined
Dec 11, 2019
Messages
286 (0.18/day)
I don't think Alder Lake's 'base power' is any different to their previous CPUs, they're also supposed to dissipate this at their base clock.

I've run a fair few Intel CPUs without turbo enabled and default power limits and I've found it to be accurate (the likes of Prime not withstanding), but compared to 10th gen (or earlier), Rocket Lake has to throttle aggressively to hit those limits, even in some games, which I'd assume is one reason why the base clocks are so low.

OEMs seem to build their system based on this number, with some puny air cooler, which theoretically offers that level of performance.

If you were planning to run a 10th gen CPU in games (or earlier), then a cooler that matches the TDP (i.e. a half-decent air cooler) would be loud, but sufficient. I still wouldn't want a 65 watt cooler on a 11700 though, even if it was only for games. If you're buying a premium AIO, then I doubt you'll run into issues when gaming with any CPU, even with no power limits enabled.
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.21/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
It's whatever they want it to mean, at the time.

I think it's marketing taking a term from manufacturing, because the TDP values are in Tier levels.
You can have multiple products in the same series, in the same launch with the same TDP value, but different eletrical and thermal values.

Why? Because then parts manufacturers need to say "This board is rated for products that average at, or below this value"
Like the rapidly vanishing "This cooler is for 65W and under CPU's only"

Motherboard VRM's, CPU Cooling, power supply requirements, etc.
Then we get products with time limits, and multiple tiers in the same product.
(12900K Compatible? 100% manufacturer supported, but only at PL1 for 6 seconds per minute per their estimates on the type of average use case the average office worker in their average study group used)


There is of course, variance and tolerances on top of that by every product involved, marketing, budget cuts, and we end up with terms that mean almost nothing in the end.

Us, as the end users want to know "What's it gunna take, for the worst case thermal and power scenario so we can plan for it?"
And TDP doesn't tell us that.
 
Joined
Apr 30, 2011
Messages
2,648 (0.56/day)
Location
Greece
Processor AMD Ryzen 5 5600@80W
Motherboard MSI B550 Tomahawk
Cooling ZALMAN CNPS9X OPTIMA
Memory 2*8GB PATRIOT PVS416G400C9K@3733MT_C16
Video Card(s) Sapphire Radeon RX 6750 XT Pulse 12GB
Storage Sandisk SSD 128GB, Kingston A2000 NVMe 1TB, Samsung F1 1TB, WD Black 10TB
Display(s) AOC 27G2U/BK IPS 144Hz
Case SHARKOON M25-W 7.1 BLACK
Audio Device(s) Realtek 7.1 onboard
Power Supply Seasonic Core GC 500W
Mouse Sharkoon SHARK Force Black
Keyboard Trust GXT280
Software Win 7 Ultimate 64bit/Win 10 pro 64bit/Manjaro Linux
For both companies, TDP is the max power in W that their CPUs are allowed to consume when the limiters are enabled. Due to motherboard vendors competition to be advertised as better than the others they have the UEFI settings altered to allow higher power draw and thus, the most of the boards allow power draw to go much higher than the TDP of the CPU installed.
 
Joined
Jul 25, 2006
Messages
12,011 (1.86/day)
Location
Nebraska, USA
System Name Brightworks Systems BWS-6 E-IV
Processor Intel Core i5-6600 @ 3.9GHz
Motherboard Gigabyte GA-Z170-HD3 Rev 1.0
Cooling Quality case, 2 x Fractal Design 140mm fans, stock CPU HSF
Memory 32GB (4 x 8GB) DDR4 3000 Corsair Vengeance
Video Card(s) EVGA GEForce GTX 1050Ti 4Gb GDDR5
Storage Samsung 850 Pro 256GB SSD, Samsung 860 Evo 500GB SSD
Display(s) Samsung S24E650BW LED x 2
Case Fractal Design Define R4
Power Supply EVGA Supernova 550W G2 Gold
Mouse Logitech M190
Keyboard Microsoft Wireless Comfort 5050
Software W10 Pro 64-bit
I think it is all hogwash. I think it started out by the engineers as a proper technical term, but then, due to advances in technologies, along with design changes that came with those advances, the manner in which TDP was determined and measured changed too. Same with temperatures.

This led to inconsistencies in how the value was determined, where sensors were physically located within and on the die, how they were measured, etc. etc.

This resulted in TDP meaning one thing with this CPU, and TDP meaning something a bit different on that CPU - even within the same brand.

You can see this in the spec sheets of the various CPUs. Some CPU report TDP, others don't. Some use T-junction, others don't. Some specify "Max. Operating Temperature (Tjmax)", others don't - just as a couple examples. And then when they do report something, do we really know what it means? For example, on some processors, Intel reports the "Processor Base Power". What's that? Simple, it is,
The time-averaged power dissipation that the processor is validated to not exceed during manufacturing while executing an Intel-specified high complexity workload at Base Frequency and at the junction temperature as specified in the Datasheet for the SKU segment and configuration.

Got it?

On top of that, because there are no industry standards on this (note in that quote it says, "Intel-specified workload",) Intel and AMD don't use the same methods and protocols for determining, sensing, measuring, and reporting those values either. So you cannot really compare a CPU from AMD with a CPU from Intel in those categories.

Then, to add to the confusion, the marketing weenies started sticking their noses and grubby fingers in the mix, putting their own "spin" on the terms. :(

So what does it really mean? (my bold underline added for emphasis)
It's whatever they want it to mean, at the time.
 
Joined
Feb 6, 2021
Messages
2,553 (2.23/day)
Location
Germany
System Name Sunk Cost Fallacy
Processor AMD Ryzen 7 7800X3D
Motherboard ASRock B650E Steel Legend Wifi
Cooling Arctic Liquid Freezer II 360 Rev. 7
Memory 2x16GB G.Skill Trident Z5 NEO 6000 CL30
Video Card(s) Sapphire Nitro+ RX 7900 XTX Vapor-X
Storage WD Black SN850X 1TB + 2x 2TB, 2x 4TB Crucial MX500, 4TB Samsung 870 Evo.
Display(s) Alienware AW2723DF, LG 27GR93U, LG 27GN950-B
Case Lian Li O11 Air Mini
Audio Device(s) Bose Companion Series 2 III, Sennheiser GSP600 and HD599 SE - Creative Soundblaster X4
Power Supply bequiet! Dark Power Pro 12 1500w Titanium
Mouse Logitech GPRO X Superlight & G502 X
Keyboard Corsair K65 RGB Mini, Razer Black Widow V3 TKL
VR HMD Oculus Rift S
Intel is easy to explain:
Thermal Design Power (TDP) represents the average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload. Refer to Datasheet for thermal solution requirements.

AMD on the other hand:
AMD’s definition of TDP will be dissected over the entirety of this piece, but we’ll start with the formula: TDP (Watts) = (tCase°C - tAmbient°C)/(HSF θca). Critically, and more on this later, we should note that each of these values changes based on the processor, so there’s no fixed set of values and each remains a variable for all processors. This means that TDP can be configured to equal other desirable numbers just by arbitrarily redefining values like tAmbient, or soft numbers like tCase. These can be manipulated for the desired TDP value.
 
Joined
Jan 20, 2019
Messages
1,243 (0.66/day)
Location
London, UK
System Name ❶ Oooh (2024) ❷ Aaaah (2021) ❸ Ahemm (2017)
Processor ❶ 5800X3D ❷ i7-9700K ❸ i7-7700K
Motherboard ❶ X570-F ❷ Z390-E ❸ Z270-E
Cooling ❶ ALFIII 360 ❷ X62 + X72 (GPU mod) ❸ X62
Memory ❶ 32-3600/16 ❷ 32-3200/16 ❸ 16-3200/16
Video Card(s) ❶ 3080 X Trio ❷ 2080TI (AIOmod) ❸ 1080TI
Storage ❶ NVME/SSD/HDD ❷ <SAME ❸ SSD/HDD
Display(s) ❶ 1440/165/IPS ❷ 1440/144/IPS ❸ 1080/144/IPS
Case ❶ BQ Silent 601 ❷ Cors 465X ❸ Frac Mesh C
Audio Device(s) ❶ HyperX C2 ❷ HyperX C2 ❸ Logi G432
Power Supply ❶ HX1200 Plat ❷ RM750X ❸ EVGA 650W G2
Mouse ❶ Logi G Pro ❷ Razer Bas V3 ❸ Logi G502
Keyboard ❶ Logi G915 TKL ❷ Anne P2 ❸ Logi G610
Benchmark Scores I have wrestled bandwidths, Tussled with voltages, Handcuffed Overclocks, Thrown Gigahertz in Jail
Some of this stuff just goes over my head.

In short, i don't really need to look at TDP specs when purchasing a CPU right?

I'm looking to purchase a CPU upgrade or new CPU+mobo soon. I'm assuming only power consumption benchmarks (with gaming loads) is the only relevant stats I should look at(?) Ideally, without compromising performance i want a quiet machine with my X62 Kraken cooler without the radiator fans running on those higher RPMs hence hoping to grab a more efficient CPU. I got LL120 fans on the rad and might fancy lowering the fan curve.
 
Joined
Feb 6, 2021
Messages
2,553 (2.23/day)
Location
Germany
System Name Sunk Cost Fallacy
Processor AMD Ryzen 7 7800X3D
Motherboard ASRock B650E Steel Legend Wifi
Cooling Arctic Liquid Freezer II 360 Rev. 7
Memory 2x16GB G.Skill Trident Z5 NEO 6000 CL30
Video Card(s) Sapphire Nitro+ RX 7900 XTX Vapor-X
Storage WD Black SN850X 1TB + 2x 2TB, 2x 4TB Crucial MX500, 4TB Samsung 870 Evo.
Display(s) Alienware AW2723DF, LG 27GR93U, LG 27GN950-B
Case Lian Li O11 Air Mini
Audio Device(s) Bose Companion Series 2 III, Sennheiser GSP600 and HD599 SE - Creative Soundblaster X4
Power Supply bequiet! Dark Power Pro 12 1500w Titanium
Mouse Logitech GPRO X Superlight & G502 X
Keyboard Corsair K65 RGB Mini, Razer Black Widow V3 TKL
VR HMD Oculus Rift S
In short, i don't really need to look at TDP specs when purchasing a CPU right?
yes
there are way more factors for cooling. the process node for example... 7nm ryzen chips run hot because of the high density (your heatsink is cold and the CPU runs at 80°C because it can not move the heat fast enough through the silicon, solder, ihs, thermal paste and cooler.

buy a decent cooler and you'll be fine. you always can check out some benchmark videos for coolers if you are not sure.
 
Joined
Dec 11, 2019
Messages
286 (0.18/day)
Some of this stuff just goes over my head.

In short, i don't really need to look at TDP specs when purchasing a CPU right?

I'm looking to purchase a CPU upgrade or new CPU+mobo soon. I'm assuming only power consumption benchmarks (with gaming loads) is the only relevant stats I should look at(?) Ideally, without compromising performance i want a quiet machine with my X62 Kraken cooler without the radiator fans running on those higher RPMs hence hoping to grab a more efficient CPU. I got LL120 fans on the rad and might fancy lowering the fan curve.

For best results, check they have a high-end GPU (like a 3080 or 3090) and a heavy CPU load (like Cyberpunk). YouTube videos (with footage) can be helpful for that, because most reviews don't have any gaming benchmarks, or only the one. Like said above though, buy a decent cooler and it's fine.
 
Joined
May 8, 2021
Messages
1,978 (1.87/day)
Location
Lithuania
System Name Shizuka
Processor Intel Core i5 10400F
Motherboard Gigabyte B460M Aorus Pro
Cooling Scythe Choten
Memory 2x8GB G.Skill Aegis 2666 MHz
Video Card(s) PowerColor Red Dragon V2 RX 580 8GB ~100 watts in Wattman
Storage 512GB WD Blue + 256GB WD Green + 4TH Toshiba X300
Display(s) BenQ BL2420PT
Case Cooler Master Silencio S400
Audio Device(s) Topping D10 + AIWA NSX-V70
Power Supply Chieftec A90 550W (GDP-550C)
Mouse Steel Series Rival 100
Keyboard Hama SL 570
Software Windows 10 Enterprise
Means nothing really. Sometimes chips are way under TDP, sometimes they are way over TDP in terms of power consumption and heat output. Intel defines TDP at base speed in stress test load, AMD is more like average power usage and heat output with unknown settings (most likely base clock speed). Always read reviews to see how chip actually performs, that's the only way to know its power consumption and heat output.
 
Joined
Jan 20, 2019
Messages
1,243 (0.66/day)
Location
London, UK
System Name ❶ Oooh (2024) ❷ Aaaah (2021) ❸ Ahemm (2017)
Processor ❶ 5800X3D ❷ i7-9700K ❸ i7-7700K
Motherboard ❶ X570-F ❷ Z390-E ❸ Z270-E
Cooling ❶ ALFIII 360 ❷ X62 + X72 (GPU mod) ❸ X62
Memory ❶ 32-3600/16 ❷ 32-3200/16 ❸ 16-3200/16
Video Card(s) ❶ 3080 X Trio ❷ 2080TI (AIOmod) ❸ 1080TI
Storage ❶ NVME/SSD/HDD ❷ <SAME ❸ SSD/HDD
Display(s) ❶ 1440/165/IPS ❷ 1440/144/IPS ❸ 1080/144/IPS
Case ❶ BQ Silent 601 ❷ Cors 465X ❸ Frac Mesh C
Audio Device(s) ❶ HyperX C2 ❷ HyperX C2 ❸ Logi G432
Power Supply ❶ HX1200 Plat ❷ RM750X ❸ EVGA 650W G2
Mouse ❶ Logi G Pro ❷ Razer Bas V3 ❸ Logi G502
Keyboard ❶ Logi G915 TKL ❷ Anne P2 ❸ Logi G610
Benchmark Scores I have wrestled bandwidths, Tussled with voltages, Handcuffed Overclocks, Thrown Gigahertz in Jail
yes
there are way more factors for cooling. the process node for example... 7nm ryzen chips run hot because of the high density (your heatsink is cold and the CPU runs at 80°C because it can not move the heat fast enough through the silicon, solder, ihs, thermal paste and cooler.

buy a decent cooler and you'll be fine. you always can check out some benchmark videos for coolers if you are not sure.

9900K with an existing X62 Kraken pump - im assuming is all good?

I know it's a K series chip but won't be overclocking other then the auto boosting/turbo features left enabled. Im going for a silent build.
 
Joined
May 8, 2021
Messages
1,978 (1.87/day)
Location
Lithuania
System Name Shizuka
Processor Intel Core i5 10400F
Motherboard Gigabyte B460M Aorus Pro
Cooling Scythe Choten
Memory 2x8GB G.Skill Aegis 2666 MHz
Video Card(s) PowerColor Red Dragon V2 RX 580 8GB ~100 watts in Wattman
Storage 512GB WD Blue + 256GB WD Green + 4TH Toshiba X300
Display(s) BenQ BL2420PT
Case Cooler Master Silencio S400
Audio Device(s) Topping D10 + AIWA NSX-V70
Power Supply Chieftec A90 550W (GDP-550C)
Mouse Steel Series Rival 100
Keyboard Hama SL 570
Software Windows 10 Enterprise
I think it is all hogwash. I think it started out by the engineers as a proper technical term, but then, due to advances in technologies, along with design changes that came with those advances, the manner in which TDP was determined and measured changed too. Same with temperatures.
I disagree. It's mostly for cutting costs, increasing profits and making chips appear better than they are. They would face very little difficulty if they actually were interested in providing proper metrics.
 
Joined
Jul 25, 2006
Messages
12,011 (1.86/day)
Location
Nebraska, USA
System Name Brightworks Systems BWS-6 E-IV
Processor Intel Core i5-6600 @ 3.9GHz
Motherboard Gigabyte GA-Z170-HD3 Rev 1.0
Cooling Quality case, 2 x Fractal Design 140mm fans, stock CPU HSF
Memory 32GB (4 x 8GB) DDR4 3000 Corsair Vengeance
Video Card(s) EVGA GEForce GTX 1050Ti 4Gb GDDR5
Storage Samsung 850 Pro 256GB SSD, Samsung 860 Evo 500GB SSD
Display(s) Samsung S24E650BW LED x 2
Case Fractal Design Define R4
Power Supply EVGA Supernova 550W G2 Gold
Mouse Logitech M190
Keyboard Microsoft Wireless Comfort 5050
Software W10 Pro 64-bit
I disagree. It's mostly for cutting costs, increasing profits and making chips appear better than they are.
:( Did you even read past my first sentence?

If you didn't take my sentence out of the context of my entire reply, you would see you agree with me.

Engineers care about costs, but not profits. And they definitely do not care about making products appear better than they are.

But "marketing weenies" do indeed care about all those thing you mentioned - and I was pretty clear about that, and how I feel about them, in the part of my reply you ignored.
 
Joined
May 8, 2021
Messages
1,978 (1.87/day)
Location
Lithuania
System Name Shizuka
Processor Intel Core i5 10400F
Motherboard Gigabyte B460M Aorus Pro
Cooling Scythe Choten
Memory 2x8GB G.Skill Aegis 2666 MHz
Video Card(s) PowerColor Red Dragon V2 RX 580 8GB ~100 watts in Wattman
Storage 512GB WD Blue + 256GB WD Green + 4TH Toshiba X300
Display(s) BenQ BL2420PT
Case Cooler Master Silencio S400
Audio Device(s) Topping D10 + AIWA NSX-V70
Power Supply Chieftec A90 550W (GDP-550C)
Mouse Steel Series Rival 100
Keyboard Hama SL 570
Software Windows 10 Enterprise
:( Did you even read past my first sentence?

If you didn't take my sentence out of the context of my entire reply, you would see you agree with me.
Not necessarily. I disagree that they can even mean one thing at brand A and mean thing B at brand B. "Thermal Dissipation Power" it clearly talks about temperature and power in this context about temperature, not energy consumption. If anyone starts to interpret it anyway they want, then term becomes bullshit.


Engineers care about costs, but not profits. And they definitely do not care about making products appear better than they are.
You never know, they might have internal fights about TDP.

But "marketing weenies" do indeed care about all those thing you mentioned - and I was pretty clear about that, and how I feel about them, in the part of my reply you ignored.
I didn't ignore it. I agree with it, but I think it's more about marketing than you said (if not 100% about marketing) and has been that way pretty much since Pentium 4/Athlon 64 era.
 
Joined
Feb 3, 2017
Messages
3,475 (1.33/day)
Processor R5 5600X
Motherboard ASUS ROG STRIX B550-I GAMING
Cooling Alpenföhn Black Ridge
Memory 2*16GB DDR4-2666 VLP @3800
Video Card(s) EVGA Geforce RTX 3080 XC3
Storage 1TB Samsung 970 Pro, 2TB Intel 660p
Display(s) ASUS PG279Q, Eizo EV2736W
Case Dan Cases A4-SFX
Power Supply Corsair SF600
Mouse Corsair Ironclaw Wireless RGB
Keyboard Corsair K60
VR HMD HTC Vive
I understand TDP has nothing to do with actual power consumption. I also understand TDP is in reference to heat output. Why its a necessary measurement, beats me (cooler related?)
TDP should define maximum heat dissipation of the chip. In practice this is the maximum power consumption of the chip or for today's CPUs - power limit. Both Intel and AMD have perverted the term. For how much heat the CPU can emit at maximum:
- Intel's TDP is meaningless, look at PL1/PL2 instead
- AMD has power limit in most cases at TDP*1.35
For both companies, TDP is the max power in W that their CPUs are allowed to consume when the limiters are enabled. Due to motherboard vendors competition to be advertised as better than the others they have the UEFI settings altered to allow higher power draw and thus, the most of the boards allow power draw to go much higher than the TDP of the CPU installed.
It should be and was exactly that for a long time. Not true any more since Intel 8000 series and Ryzen 3000 series. Motherboard vendors are a scapegoat.

Motherboard manufacturers' trickery was kind of true for Intel 8000 series but after that all went to hell. Intel did not enforce any of the spec limits and encouraged motherboard manufacturers to exceed or remove limits. AMD simply went for the more complex TDP definition and power limit higher than TDP.
along with design changes that came with those advances, the manner in which TDP was determined and measured changed too. Same with temperatures.
This led to inconsistencies in how the value was determined, where sensors were physically located within and on the die, how they were measured, etc. etc.
Nope. Considering how and how well CPUs manage both voltage and current fed into them the power measurement is definitely exact enough.
"Thermal Dissipation Power" it clearly talks about temperature and power in this context about temperature, not energy consumption. If anyone starts to interpret it anyway they want, then term becomes bullshit.
For a CPU, the power that goes in comes out as heat. Any other sink for energy is very-very-very negligible.
 
Last edited:
Joined
Dec 11, 2019
Messages
286 (0.18/day)
9900K with an existing X62 Kraken pump - im assuming is all good?

I know it's a K series chip but won't be overclocking other then the auto boosting/turbo features left enabled. Im going for a silent build.

From this video, it seems to stay under, or near to the 95 watt tdp when at stock settings:

But, overclocked to 5.1 Ghz, in some games, it can reach up to 150 watt:

I think you'd be fine at stock, but some voltage tuning could help if it's too hot for your liking.

I'm not sure how hard you'd have to work the cooler though, so I can't say if it will be quiet. If you turned off the turbo it would help, though it'll also give you way less performance. V-sync can be similar if you're fine with 60 fps.
 

Mussels

Freshwater Moderator
Staff member
Joined
Oct 6, 2004
Messages
58,413 (8.21/day)
Location
Oystralia
System Name Rainbow Sparkles (Power efficient, <350W gaming load)
Processor Ryzen R7 5800x3D (Undervolted, 4.45GHz all core)
Motherboard Asus x570-F (BIOS Modded)
Cooling Alphacool Apex UV - Alphacool Eisblock XPX Aurora + EK Quantum ARGB 3090 w/ active backplate
Memory 2x32GB DDR4 3600 Corsair Vengeance RGB @3866 C18-22-22-22-42 TRFC704 (1.4V Hynix MJR - SoC 1.15V)
Video Card(s) Galax RTX 3090 SG 24GB: Underclocked to 1700Mhz 0.750v (375W down to 250W))
Storage 2TB WD SN850 NVME + 1TB Sasmsung 970 Pro NVME + 1TB Intel 6000P NVME USB 3.2
Display(s) Phillips 32 32M1N5800A (4k144), LG 32" (4K60) | Gigabyte G32QC (2k165) | Phillips 328m6fjrmb (2K144)
Case Fractal Design R6
Audio Device(s) Logitech G560 | Corsair Void pro RGB |Blue Yeti mic
Power Supply Fractal Ion+ 2 860W (Platinum) (This thing is God-tier. Silent and TINY)
Mouse Logitech G Pro wireless + Steelseries Prisma XL
Keyboard Razer Huntsman TE ( Sexy white keycaps)
VR HMD Oculus Rift S + Quest 2
Software Windows 11 pro x64 (Yes, it's genuinely a good OS) OpenRGB - ditch the branded bloatware!
Benchmark Scores Nyooom.
The problem is that 'high complexity workloads' can mean anything from an excel spreadsheet to AVX-512, and the power consumption and heat output vary widly.


The numbers give us a starting point, and then we just get power and cooling bigger than the misleading number they give us. Easy!
 
Joined
Jul 25, 2006
Messages
12,011 (1.86/day)
Location
Nebraska, USA
System Name Brightworks Systems BWS-6 E-IV
Processor Intel Core i5-6600 @ 3.9GHz
Motherboard Gigabyte GA-Z170-HD3 Rev 1.0
Cooling Quality case, 2 x Fractal Design 140mm fans, stock CPU HSF
Memory 32GB (4 x 8GB) DDR4 3000 Corsair Vengeance
Video Card(s) EVGA GEForce GTX 1050Ti 4Gb GDDR5
Storage Samsung 850 Pro 256GB SSD, Samsung 860 Evo 500GB SSD
Display(s) Samsung S24E650BW LED x 2
Case Fractal Design Define R4
Power Supply EVGA Supernova 550W G2 Gold
Mouse Logitech M190
Keyboard Microsoft Wireless Comfort 5050
Software W10 Pro 64-bit
You never know, they might have internal fights about TDP.
LOL

Clearly, you are not a design engineer. If engineers on a design project are worried about "profits", then they are not really on the design team to design. Unless the engineer has a stake in the company itself, engineers care about costs, staying within budgets, and avoiding cost overruns. Profits are the jobs of marketing and execs. Engineers care about the most bang for the money, not the most money for the bang.
 
Joined
May 8, 2021
Messages
1,978 (1.87/day)
Location
Lithuania
System Name Shizuka
Processor Intel Core i5 10400F
Motherboard Gigabyte B460M Aorus Pro
Cooling Scythe Choten
Memory 2x8GB G.Skill Aegis 2666 MHz
Video Card(s) PowerColor Red Dragon V2 RX 580 8GB ~100 watts in Wattman
Storage 512GB WD Blue + 256GB WD Green + 4TH Toshiba X300
Display(s) BenQ BL2420PT
Case Cooler Master Silencio S400
Audio Device(s) Topping D10 + AIWA NSX-V70
Power Supply Chieftec A90 550W (GDP-550C)
Mouse Steel Series Rival 100
Keyboard Hama SL 570
Software Windows 10 Enterprise
LOL

Clearly, you are not a design engineer. If engineers on a design project are worried about "profits", then they are not really on the design team to design. Unless the engineer has a stake in the company itself, engineers care about costs, staying within budgets, and avoiding cost overruns. Profits are the jobs of marketing and execs. Engineers care about the most bang for the money, not the most money for the bang.
Depends on engineer. Some want to make the best they can, some want something simple, some want to pinch pennies, some are just happy to have a decent job.
 
Top