azraelI guess BFC means the same thing as BFG, just with 'cache' replacing 'gun'. IYKYK. :p

Yeah that was my assumption too, and I really like it.

I mean they'll probably have another official meaning, but I'm 100% certain it means Big F*ing Cache and that it's a nod to Doom/Quake.

bugYes, real easy. Now add in heterogeneous cores, low power versions/various TDP targets, different number of IGP cores. Gets pretty trivial real fast.

Why do you need to add them in the model name? There isn't even 10 models in any given generation of desktop CPUs.

BlueScreenJunkyYeah that was my assumption too, and I really like it.

I mean they'll probably have another official meaning, but I'm 100% certain it means Big F*ing Cache and that it's a nod to Doom/Quake.

To me, it looks more like JFC for some common swearing. Not to spoil the fun, but I do think such abbreviations in an actual product name are childish. Each to their own.

DrUsmanSaleemwas reportedly set to pull up to 296 W at PL2 despite its 125 W rating

I really don't understand then. Why put 125W as the rating, if you'll need a lot better cooler to properly handle multicore workloads?
Seems like marketing just to compare to the competition.

Also, why are some people already saying here they'll replace their hardware, if you havent seen any benchmarks yet?

Trying to understand the hype: is it the cache or the cores? Whenever I see cpus with such power draw I get an "oh well", becase they seem rather inefficient. Of course time will tell and hopefully these bring great competition against AMD :)

JyubeiI really don't understand then. Why put 125W as the rating, if you'll need a lot better cooler to properly handle multicore workloads?
Seems like marketing just to compare to the competition.

Cause a 125w cooler will handle 250w on a big intel chip just fine.

As you can see from the review, a tiny single tower 30$ cooler can handle 285 watts

www.techpowerup.com/review/fsp-np5-black-cpu-air-cooler/8.html

JyubeiAlso, why are some people already saying here they'll replace their hardware, if you havent seen any benchmarks yet?

Trying to understand the hype: is it the cache or the cores? Whenever I see cpus with such power draw I get an "oh well", becase they seem rather inefficient. Of course time will tell and hopefully these bring great competition against AMD :)

Power draw is the most irrelevant thing on a cpu cause it's one of the few things that is completely up to you to change. Not only does it takes 5 seconds to change the power limit, you can also buy a non k sku that comes with lower power limits from the factory.

It has nothing to do with efficiency. In order to properly compare efficiency you need to put 2 Cpus at the same power level and see which one is faster.

JustBenchingCause a 125w cooler will handle 250w on a big intel chip just fine.

As you can see from the review, a tiny single tower 30$ cooler can handle 285 watts

www.techpowerup.com/review/fsp-np5-black-cpu-air-cooler/8.html

The "just fine" is a bit of a stretch.
The 285 W is right at the throttle point (100°C) and at max rpm.
For noise normalized 250 W corresponds to 92.6°C.
This is in the following conditions:

Ambient temperatures are the same as the AMD test rig at 20-21°C.

And in this type of "case":
tpucdn.com/review/fsp-np5-black-cpu-air-cooler/images/installed2.jpg

Good luck getting the same temps at 24°C ambient and in a regular mid-tower case.

Sol_BadguyThe "just fine" is a bit of a stretch.
The 285 W is right at the throttle point (100°C) and at max rpm.
For noise normalized 250 W corresponds to 92.6°C.
This is in the following conditions:

And in this type of "case":
tpucdn.com/review/fsp-np5-black-cpu-air-cooler/images/installed2.jpg

Good luck getting the same temps at 24°C ambient and in a regular mid-tower case.

It's one of the if not the cheapest cooler tested man. You are arguing just for arguing.

AusWolfWhy do you need to add them in the model name?

Gee, Idk, maybe because any of those varies, we're talking about a different CPU?

AusWolfThere isn't even 10 models in any given generation of desktop CPUs.

Not even 10:
en.wikipedia.org/wiki/Arrow_Lake_(microprocessor)#Desktop
en.wikipedia.org/wiki/Zen_5#Desktop

bugGee, Idk, maybe because any of those varies, we're talking about a different CPU?

Not even 10:
en.wikipedia.org/wiki/Arrow_Lake_(microprocessor)#Desktop
en.wikipedia.org/wiki/Zen_5#Desktop

Point taken. Still, it doesn't explain why you need a hundred 0s and an X on AMD, or why every single model has to end in a "5" on Intel. If you take that out, you have 2 digits and a letter. More than enough, imo.

JyubeiI really don't understand then. Why put 125W as the rating, if you'll need a lot better cooler to properly handle multicore workloads?

If anything, high load on all cores are far easier to cool than a few hot spots.

Sol_BadguyThe "just fine" is a bit of a stretch.
The 285 W is right at the throttle point (100°C) and at max rpm.

Most decent tower coolers can handle >250W just fine in theory, but in practice it depends on the hot spots on the die. This is where the differences in cooler design and CPU models come in.

But there's no reason to go for a 280mm AIO unless it's for heavy overlocking, as a tower cooler with more heat pipes will do just fine, if not better in practice. My go-to option is the Noctua NH-U14S, which should be fine for any of these models, and likely upcoming Nova Lake and Zen 6 too, unless they have even worse thermal distribution. But there are many cheaper options too. For reference, the workstation version of NH-U14S tops out at 680W, so the limiting factor here is the heat transfer between the die and the cooler.

Sol_BadguyAnd in this type of "case":
tpucdn.com/review/fsp-np5-black-cpu-air-cooler/images/installed2.jpg

Which will only skew the real world results in favor of tower coolers;
- Realistically, very few cases have air flow to accommodate a 280+mm AIO at a capacity where it beat a good tower cooler.
- Buying a high-performing tower cooler will leave room in the budget to upgrade the case fans, which will be the limiting factor anyways. For sustained loads the case cooling should exceed the thermal capacity of the CPU cooler, otherwise you're just wasting money. Even a $30 tower cooler with the right case cooling will be wiser than a hefty AIO with low case airflow (which most buyers have).
- Having a good setup with case fans will even help the efficiency of the tower cooler vs. an open test rig.

efikkanIf anything, high load on all cores are far easier to cool than a few hot spots.

That's why loading 5-6 cores on my 9700X makes it run hotter than loading all 8. Also, when the power budget is distributed among fewer cores, they will eat more power per core, and thus, run hotter. That's why, when one's testing for cooling, ST and MT loads aren't enough. You need a variety of scenarios.

Anyway, I digress.

WirkoIntel's intel service collected some intel indicating that AMD is releasing five-digit processors soon. You can't fight that with three-digit processors, much less if they lack the "X" at the end.

If you'll recall, AMD switched to three-digit desktop processor naming at the beginning of this year with their paper-launched 400-series Gorgon Point Desktop.

Intel's naming scheme changes are usually done to differentiate from AMD, who commonly follow in Intel's footsteps.

JustBenchingCause a 125w cooler will handle 250w on a big intel chip just fine.

As you can see from the review, a tiny single tower 30$ cooler can handle 285 watts

www.techpowerup.com/review/fsp-np5-black-cpu-air-cooler/8.html


Power draw is the most irrelevant thing on a cpu cause it's one of the few things that is completely up to you to change. Not only does it takes 5 seconds to change the power limit, you can also buy a non k sku that comes with lower power limits from the factory.

It has nothing to do with efficiency. In order to properly compare efficiency you need to put 2 Cpus at the same power level and see which one is faster.

This is a really great answer to all the people who complain about power draw. Intel is not forcing you to run your cpu at the power limits all day long to get great performance. Overclocking now does not give anywhere near the performance increase it did before without getting ridicoulous cooling and we are getting closer to the limits of physic for all these cpu's every generation. If you have no intention of getting the top model, why do you care about max power rating and its power consumption!!

interested in BLLC performance vs non-BLLC on the ultra 9s.

JustBenchingPeople dreaming zen6 beating nova lake when in reality watt for watt it won't even beat arrow lake. Mark my words, the 200 and 300$ zen 6 offerings are going to get pummeled by the 250k / 270k at iso watt.

Watt for watt arrow lake loses go zen5, youve gotten more and mkre extreme even thouh ive disproved this basless claim time and time again. Intels cpus are muvh less efficient, needing better nodes to get close to tge same tdp. Arrow lake would draw raptor lake power if it was on n4. The 180$ apparently 8 core silver king chip gets 8 zen6 cores which shou!d match 10 zen5 cores. Its a 7400f competitor

JustBenchingCause a 125w cooler will handle 250w on a big intel chip just fine.

As you can see from the review, a tiny single tower 30$ cooler can handle 285 watts

www.techpowerup.com/review/fsp-np5-black-cpu-air-cooler/8.html


Power draw is the most irrelevant thing on a cpu cause it's one of the few things that is completely up to you to change. Not only does it takes 5 seconds to change the power limit, you can also buy a non k sku that comes with lower power limits from the factory.

It has nothing to do with efficiency. In order to properly compare efficiency you need to put 2 Cpus at the same power level and see which one is faster.

yes and the 9950x beats or matches a 290w arrow lake chip at 210w you dont need to compare at the same W. Amd dosent have offerings that guzzle 290w. Power limmiting can benifit the chips on better nodes, as they take less voltage to reach the same clocks. If a chip at lower wattage beats or matches a higher wattage part(despite being on a 8% more efficient node) than the first cpu is vastly more efficient.

JustBenchingCause a 125w cooler will handle 250w on a big intel chip just fine.

As you can see from the review, a tiny single tower 30$ cooler can handle 285 watts

www.techpowerup.com/review/fsp-np5-black-cpu-air-cooler/8.html


Power draw is the most irrelevant thing on a cpu cause it's one of the few things that is completely up to you to change. Not only does it takes 5 seconds to change the power limit, you can also buy a non k sku that comes with lower power limits from the factory.

It has nothing to do with efficiency. In order to properly compare efficiency you need to put 2 Cpus at the same power level and see which one is faster.

Damm this bullshit again. Im sorry, does mobile not matter to you? Do people in countries where power is extremly expensive have to fork a loan out? Do we need space heaters again? You switch your stance or say "oh but buy a locked t chip with shitty multicore performance or power limmit and loose a shit ton of performance to match amd's powrr draw" when you are arguing this on other forums the exact opposite. You keep moving the damm goalpost. Could have taklked about arrow lakes value but nooooooo you have to argue a point that doesn't line up with reality.
Power draw does fucking matter. Arrow lake at 210w doesn't perform like amd at 210w if you get rid of thier node advatgae with math. It actually performs like shit comparatively. You loose quite a bit of multithread if you power limmit arrow lake, where it barley matches the 9950x before it now looses by quite a bit.

tsurugi

Amd chiplets packed closer together, intel got e cores which are more spread out, and power draw too. 293 vs 220w, and intrl shaved its dies and interfaces as close as possible going all the way back to 10th gen, so this is unsuprising.

randomdude9182Watt for watt arrow lake loses go zen5, youve gotten more and mkre extreme even thouh ive disproved this basless claim time and time again. Intels cpus are muvh less efficient, needing better nodes to get close to tge same tdp. Arrow lake would draw raptor lake power if it was on n4. The 180$ apparently 8 core silver king chip gets 8 zen6 cores which shou!d match 10 zen5 cores. Its a 7400f competitor


yes and the 9950x beats or matches a 290w arrow lake chip at 210w you dont need to compare at the same W. Amd dosent have offerings that guzzle 290w. Power limmiting can benifit the chips on better nodes, as they take less voltage to reach the same clocks. If a chip at lower wattage beats or matches a higher wattage part(despite being on a 8% more efficient node) than the first cpu is vastly more efficient.


Damm this bullshit again. Im sorry, does mobile not matter to you? Do people in countries where power is extremly expensive have to fork a loan out? Do we need space heaters again? You switch your stance or say "oh but buy a locked t chip with shitty multicore performance or power limmit and loose a shit ton of performance to match amd's powrr draw" when you are arguing this on other forums the exact opposite. You keep moving the damm goalpost. Could have taklked about arrow lakes value but nooooooo you have to argue a point that doesn't line up with reality.
Power draw does fucking matter. Arrow lake at 210w doesn't perform like amd at 210w if you get rid of thier node advatgae with math. It actually performs like shit comparatively. You loose quite a bit of multithread if you power limmit arrow lake, where it barley matches the 9950x before it now looses by quite a bit.


Amd chiplets packed closer together, intel got e cores which are more spread out, and power draw too. 293 vs 220w, and intrl shaved its dies and interfaces as close as possible going all the way back to 10th gen, so this is unsuprising.

Can you show me a 200 or 300$ AMD chip beating the 250 or the 270k at the same wattage in any MT workload ?

If not, you are wrong, move on.

randomdude9182Amd chiplets packed closer together, intel got e cores which are more spread out, and power draw too. 293 vs 220w, and intrl shaved its dies and interfaces as close as possible going all the way back to 10th gen, so this is unsuprising.

R23 Efficiency
270k plus > 285k > 9950x

CPU Temperature
9950x >> 285k > 270k plus

Node performance
N3B <= N4P

N3B +10% vs N5
N4P +11% vs N5

AMD uses SMT while Intel does not, and SMT is extremely beneficial for improving MT efficiency, even in this case, where Intel and AMD use nodes with similar performance, the efficiency of the 9950x is lower than that of the 285k.

I expect that for novalake and zen6, the MT performance and efficiency of the Ultra 5 4650K will be much higher than a full 12 core Ryzen 7.

tsurugi

R23 Efficiency
270k plus > 285k > 9950x

CPU Temperature
9950x >> 285k > 270k plus

Node performance
N3B <= N4P

N3B +10% vs N5
N4P +11% vs N5

AMD uses SMT while Intel does not, and SMT is extremely beneficial for improving MT efficiency, even in this case, where Intel and AMD use nodes with similar performance, the efficiency of the 9950x is lower than that of the 285k.

I expect that for novalake and zen6, the MT performance and efficiency of the Ultra 5 4650K will be much higher than a full 12 core Ryzen 7.

As expected, the 270k is beating the socks off of the similarly priced 9700x. Heck it even beats the twice as expensive 9950x. That's why I can't see zen 6 beating arrow lake, let alone nova lake. The gap is stupid large in efficiency.

If I understand your graph correctly, a 270k at 35w matches the 9700x at 160 watts. So the 9700x needs 500% more power.... But people are trying to convince me amd is efficient. Oh well

randomdude9182Damm this bullshit again. Im sorry, does mobile not matter to you? Do people in countries where power is extremly expensive have to fork a loan out? Do we need space heaters again? You switch your stance or say "oh but buy a locked t chip with shitty multicore performance or power limmit and loose a shit ton of performance to match amd's powrr draw" when you are arguing this on other forums the exact opposite. You keep moving the damm goalpost. Could have taklked about arrow lakes value but nooooooo you have to argue a point that doesn't line up with reality.
Power draw does fucking matter. Arrow lake at 210w doesn't perform like amd at 210w if you get rid of thier node advatgae with math. It actually performs like shit comparatively. You loose quite a bit of multithread if you power limmit arrow lake, where it barley matches the 9950x before it now looses by quite a bit.

Using AMD will consume more power. As can be seen from TPU's testing, Intel's wall socket power consumption is lower. When gaming, Intel CPU package power consumption is higher, but if you subtract the difference in wall socket power consumption, it is about the same as AMD or even more power-efficient.
For an Intel CPU with much higher MT performance, consuming about the same amount of power in gaming as AMD shows that AMD has not controlled chipset power consumption well.

randomdude9182Watt for watt arrow lake loses go zen5, youve gotten more and mkre extreme even thouh ive disproved this basless claim time and time again.

From multiple independent testing you would know its false and only representative in gaming which is the loud minority.

JustBenchingIf I understand your graph correctly, a 270k at 35w matches the 9700x at 160 watts. So the 9700x needs 500% more power.... But people are trying to convince me amd is efficient. Oh well

Not quite.
At 18000 points:
- the 270K draws 35 W
- the 285K about 38 W
- 265K and 250K about 42 W
- the 245K draws 50 W
- the 9700X about 87-88 W
- and the 9600X can't reach 18000 even at 130 W

Sol_BadguyNot quite.
At 18000 points:
- the 270K draws 35 W
- the 285K about 38 W
- 265K and 250K about 42 W
- the 245K draws 50 W
- the 9700X about 87-88 W
- and the 9600X can't reach 18000 even at 130 W

You are right, I'm on the phone and it's hard to look at the graph. Was looking at the 24k line. It's 50w for the 270k and probably 200 for the 9700x / 9800x3d. So not 500%, it's only 400%. Still quite a large gap.

Was even fx that far behind?

JustBenchingYou are right, I'm on the phone and it's hard to look at the graph. Was looking at the 24k line. It's 50w for the 270k and probably 200 for the 9700x / 9800x3d. So not 500%, it's only 400%. Still quite a large gap.

Was even fx that far behind?

One does not simply buy a 9700X for MT performance / productivity, that's my conclusion. It's 8 cores vs erm... how many on Intel? I lost count. :D

The top graph with the 9950X makes more sense for this kind of comparison. But even then, AMD had better step up the game with Zen 6, as the 270K seems to be a winner here, too (it would have been nice to see the two in the same graph, but it's fine).

AusWolfOne does not simply buy a 9700X for MT performance / productivity, that's my conclusion.

Yes, because it's slow and inefficient in MT for the price. That's not really an excuse.

Imagine Nvidia introduced a 4gb VRAM GPU for 1k$ and I said" one does not buy a 4gb vram GPU for games". Well yeah, cause it's crap. It shouldn't have 4gb or it should be considerably cheaper.

But to entertain the thought, what does someone buy the 9700x for? It's inferior to everything compared to other products in its price range, is it not?

JustBenchingYes, because it's slow and inefficient in MT for the price. That's not really an excuse.

Imagine Nvidia introduced a 4gb VRAM GPU for 1k$ and I said" one does not buy a 4gb vram GPU for games". Well yeah, cause it's crap. It shouldn't have 4gb or it should be considerably cheaper.

Not a really good analogy because...

JustBenchingBut to entertain the thought, what does someone buy the 9700x for? It's inferior to everything compared to other products in its price range, is it not?

Games. If you only game on your PC, then MT in Cinebench is irrelevant.

But you know what's worse than an 8-core CPU for £250? An 8-core CPU for £450, the 9800 and 9850X3D. Nobody seems to be roasting those two for some reason, although I think they would deserve it more than the 9700X does.