Review Intel Core Ultra 9 285K Review: Intel Throws a Lateral with Arrow Lake

Page 10 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

TheHerald

Respectable
BANNED
Feb 15, 2024
1,633
501
2,060
A content creators perspective and why intel has 80+% of marketshare. Besides the inherit unfairness of the test since the i5 parts are much cheaper and lower tier, the differences in performance are staggering. But hey, Intel bad

untitled.jpg

aaa.jpg

07-A1-A476-A6-D6-4-F28-B747-C72-FCB1-F4-F37.png
 

YSCCC

Commendable
Dec 10, 2022
566
460
1,260
Nice try chief,

View: https://youtu.be/vjPXOurg0nU?si=f82h1teMtC6Boczd


The same content creator perspective for the 285k, which is what a content creator should go for, listen for the conclusion: if you are doing photo editing, go for AMD ( and that’s only comparing 9900X), if you are doing video editing, maybe go for 285k, if you are on 12th gen, go for 14 th gen even for video as it is faster and cheaper.

Yea good release, really great product with zen 5 being total crap
 

TheHerald

Respectable
BANNED
Feb 15, 2024
1,633
501
2,060
Nice try chief,

View: https://youtu.be/vjPXOurg0nU?si=f82h1teMtC6Boczd


The same content creator perspective for the 285k, which is what a content creator should go for, listen for the conclusion: if you are doing photo editing, go for AMD ( and that’s only comparing 9900X), if you are doing video editing, maybe go for 285k, if you are on 12th gen, go for 14 th gen even for video as it is faster and cheaper.

Yea good release, really great product with zen 5 being total crap
Yeah, he said unless you are doing photo editing, get Intel. :eek:
 

YSCCC

Commendable
Dec 10, 2022
566
460
1,260
Yeah, he said unless you are doing photo editing, get Intel. :eek:
lol nice try of twisting the message.

He literally said that unless you are only doing video transcoding mostly, which the new igpu have hardware acceleration, you MIGHT want to consider the 285k, if you do gaming or photo editing, go AMD or even last gen Intel.

In strict language: except you will want the new igpu acceleration, dont buy the 285k.

Worse still it’s only comparing to the 9900X, not 9950X. But yea, Intel is the best as always.
 

TheHerald

Respectable
BANNED
Feb 15, 2024
1,633
501
2,060
lol nice try of twisting the message.

He literally said that unless you are only doing video transcoding mostly, which the new igpu have hardware acceleration, you MIGHT want to consider the 285k, if you do gaming or photo editing, go AMD or even last gen Intel.

In strict language: except you will want the new igpu acceleration, dont buy the 285k.

Worse still it’s only comparing to the 9900X, not 9950X. But yea, Intel is the best as always.
Man, come on, He said stick to 14th gen if you are already on the platform, which is kinda obvious, if you have a 12th gen it's much cheaper to go to 14 than the 285k.
 

bioshark

Distinguished
Nov 26, 2015
7
1
18,515
I have just finished (last week) upgrading my wife's gaming machine using intel i7 14700k.
I was waiting for the 15th gen (aka the new Intel Core Ultra 7 265K) to come out to compare the performance so that I may upgrade my own gaming machine early next year. Seeing all the bad reviews regarding gaming performance regression, I'll stick to i7 14700k for my machine as well, so it's time to hunt some discounts.
 
I have just finished (last week) upgrading my wife's gaming machine using intel i7 14700k.
I was waiting for the 15th gen (aka the new Intel Core Ultra 7 265K) to come out to compare the performance so that I may upgrade my own gaming machine early next year. Seeing all the bad reviews regarding gaming performance regression, I'll stick to i7 14700k for my machine as well, so it's time to hunt some discounts.
If you are just doing gaming than the 7800X3D is currently the best overall gaming CPU available. However, depending on the GPU being used the differences between the CPUs will be closer.
 

bioshark

Distinguished
Nov 26, 2015
7
1
18,515
If you are just doing gaming than the 7800X3D is currently the best overall gaming CPU available. However, depending on the GPU being used the differences between the CPUs will be closer.
I am also doing coding on my own machine, so it's not just for gaming. As for video cards, we never go crazy and stick to the XX70 series as a balance between price and performance. For the machine I built last week I bought RTX 4070 and for myself I plan to buy the same, maybe the Super edition.
 
  • Like
Reactions: jeremyj_83

YSCCC

Commendable
Dec 10, 2022
566
460
1,260
I am also doing coding on my own machine, so it's not just for gaming. As for video cards, we never go crazy and stick to the XX70 series as a balance between price and performance. For the machine I built last week I bought RTX 4070 and for myself I plan to buy the same, maybe the Super edition.
then that should be roughly the balance point, if you arn't worried about RPL power draw and hopefully the degradation is really fixed, it could be done if a deal is there, though if you're not already on a 1700 board, AM5 might make more sense? likely you could have another swarp later for zen 6 or even 7
 
A content creators perspective and why intel has 80+% of marketshare. Besides the inherit unfairness of the test since the i5 parts are much cheaper and lower tier, the differences in performance are staggering. But hey, Intel bad

untitled.jpg

aaa.jpg

07-A1-A476-A6-D6-4-F28-B747-C72-FCB1-F4-F37.png
ahhhh... this is the old Bulldozer argument writ large.

there was even a big content creator who made a series of videos arguing you were better off getting Bulldozer/piledriver then sandy bridge if you wanted to stream gameplay. arguing AMD's gaming performance was "good enough".

that's basically this argument. all over again. only being made from team blue's side.

this truly is the bulldozer of this generation.
 

TheHerald

Respectable
BANNED
Feb 15, 2024
1,633
501
2,060
ahhhh... this is the old Bulldozer argument writ large.

there was even a big content creator who made a series of videos arguing you were better off getting Bulldozer/piledriver then sandy bridge if you wanted to stream gameplay. arguing AMD's gaming performance was "good enough".

that's basically this argument. all over again. only being made from team blue's side.

this truly is the bulldozer of this generation.
Was bulldozer as fast as the fastest Intel chip at the time while consuming same or less power? Cause the 285k is as fast as the 9950x while consuming equal or less power. Stop the cope man.
 
Was bulldozer as fast as the fastest Intel chip at the time while consuming same or less power? Cause the 285k is as fast as the 9950x while consuming equal or less power. Stop the cope man.
no. you are way off on power usage... the 9950x draws literally 1/2 the power of the 285.

the 285 is "matching" the multicore performance of the 9950x with 24 actual cores vs 16 actual cores. granted there are a pile of E cores and the E core improvement for intel this cycle is truly impressive definitely the best thing about this launch.

furthermore in gaming the 9950x is slightly better on the whole then the 285, and thats not even counting the advantage of the x3D chips.


as for your question, piledriver fx8350 drew a lot more power then the i5 2500k, not quite double but it was definitely a significant jump over it. and in certain use cases, sure you could perform about as well on a 8350 as you could a 2500k. basically if you played games in 1440p and streamed at the same time you sorta had created a use environment the 8350 could match the 2500k. but at no time in it's life was piledriver a better chip then intel's sandy bridge or later advances. and an i7 would slaughter that 8350 in the aforementioned streaming scenario.
 
really now?
I'm all for discourse, but absolutely just stop with the lies.
oh for the love of...

if tech power up and a number of reviewers got the power draw wrong. it's not their fault, i'm not saying they're bad sites. intel shifted where the chip draws power from the 8 pin cpu to the 24 pin main motherboard power. most reviewers are used to measuring cpu power draw from the 8 pin. if you measure power from the 8 pin, yes you get those results.

it was the famous german overclocker der8auer who discovered the cpu was pulling power from the motherboard (and not an insignificant amount either). there s no real way to measure cpu power draw from the 24 pin main board plug without basically manually creating your own equipment to do it. der8auer and a few other tech sites he informed were able to test the actual power draw of those chips, and yes, there is a significant reduction in power draw with arrow lake. approximately 33% reduction from the prior 14th gen. but it still draws about 2x the power ryzen 5 does.

again. this isn't me banging on tech powerup, or criticizing them. it if weren't for de8auer no one else would have known. i'm sure in a few months all the tech sites will have rerun the power numbers for arrow lake and be in agreement.

as a sidenote intel isn't being deceptive or intentionally hiding the power draw. frankly intel never intended for arrow lake's main selling point to be energy efficiency. blame marketing for that.

what happened was intel was forced to drop hyperthreading from arrow lake because tsmc 3nm couldn't do hyperthreading. this resulted in some energy savings, which intel meant to push back into the chip to upclock it to hopefully brute force their way around the huge memory latency problems the arrow lake design had. however the 13th and 14th gen problem happened and intel panicked. and about 4 months ago decided to downclock arrow lake reducing the power draw so the chips won't burn themselves out like 13th and 14th gen were doing. now that intel couldn't brute force their way through the memory latency issues they needed a new selling point. and the obvious one was energy efficiency. by going from intel 10nm down to tsmc 3nm they already got some efficiency gains, combined with downclocking arrow lake and the removal of hyperthreading they could show significant power reduction in this chip.

intel didn't set out to make an efficient chip, events sort of forced them into making that the main sellng point.
 
Last edited:
  • Like
Reactions: SunMaster

TheHerald

Respectable
BANNED
Feb 15, 2024
1,633
501
2,060
no. you are way off on power usage... the 9950x draws literally 1/2 the power of the 285.

the 285 is "matching" the multicore performance of the 9950x with 24 actual cores vs 16 actual cores. granted there are a pile of E cores and the E core improvement for intel this cycle is truly impressive definitely the best thing about this launch.

furthermore in gaming the 9950x is slightly better on the whole then the 285, and thats not even counting the advantage of the x3D chips.


as for your question, piledriver fx8350 drew a lot more power then the i5 2500k, not quite double but it was definitely a significant jump over it. and in certain use cases, sure you could perform about as well on a 8350 as you could a 2500k. basically if you played games in 1440p and streamed at the same time you sorta had created a use environment the 8350 could match the 2500k. but at no time in it's life was piledriver a better chip then intel's sandy bridge or later advances. and an i7 would slaughter that 8350 in the aforementioned streaming scenario.
According to every single review I've seen the 285k draws less power on average on both workloads and gaming. In fact according to Tom's hardware the 9950x draws 40% more power on average in games for the same performance.

Even hwbusters, the guy that is THE expert on psus, and anything that has to do with it (cybernetics) said that the 9950x needs a lot more juice.

The number of cores is irrelevant. Who's fault is it that the 9950x only has 16 cores? Why would I even care?
 

TheHerald

Respectable
BANNED
Feb 15, 2024
1,633
501
2,060
oh for the love of...

if tech power up and a number of reviewers got the power draw wrong. it's not their fault, i'm not saying they're bad sites. intel shifted where the chip draws power from the 8 pin cpu to the 24 pin main motherboard power. most reviewers are used to measuring cpu power draw from the 8 pin. if you measure power from the 8 pin, yes you get those results.

it was the famous german overclocker der8auer who discovered the cpu was pulling power from the motherboard (and not an insignificant amount either). there s no real way to measure cpu power draw from the 24 pin main board plug without basically manually creating your own equipment to do it. der8auer and a few other tech sites he informed were able to test the actual power draw of those chips, and yes, there is a significant reduction in power draw with arrow lake. approximately 33% reduction from the prior 14th gen. but it still draws about 2x the power ryzen 5 does.

again. this isn't me banging on tech powerup, or criticizing them. it if weren't for de8auer no one else would have known. i'm sure in a few months all the tech sites will have rerun the power numbers for arrow lake and be in agreement.

as a sidenote intel isn't being deceptive or intentionally hiding the power draw. frankly intel never intended for arrow lake's main selling point to be energy efficiency. blame marketing for that.

what happened was intel was forced to drop hyperthreading from arrow lake because tsmc 3nm couldn't do hyperthreading. this resulted in some energy savings, which intel meant to push back into the chip to upclock it to hopefully brute force their way around the huge memory latency problems the arrow lake design had. however the 13th and 14th gen problem happened and intel panicked. and about 4 months ago decided to downclock arrow lake reducing the power draw so the chips won't burn themselves out like 13th and 14th gen were doing. now that intel couldn't brute force their way through the memory latency issues they needed a new selling point. and the obvious one was energy efficiency. by going from intel 10nm down to tsmc 3nm they already got some efficiency gains, combined with downclocking arrow lake and the removal of hyperthreading they could show significant power reduction in this chip.

intel didn't set out to make an efficient chip, events sort of forced them into making that the main sellng point.
Uhm, the cpu pulling power from the mobo only happens on a few of asus motherboards. Nothing to do with the cpu. TPU measured power with an MSI motherboard exactly for that reason

Please, let's stop the amd defense.
 
Oh how we laughed.

For months you have been defending the indefensible….
I am not going to co-sign everything @TheHerald say's because I have had my own issues with how and what they argue, but you are incorrect on this one my dude. The 285k as compared to a 9950X is; on average more power efficient, with similar enough performance in most production tasks and gaming. You can certainly get into the weeds about the specific types of power efficiency one is better at than the other, but they are very comparable on the whole of power efficiency with a slight edge to Intel.
 
Mar 10, 2020
414
376
5,070
I am not going to co-sign everything @TheHerald say's because I have had my own issues with how and what they argue, but you are incorrect on this one my dude. The 285k as compared to a 9950X is; on average more power efficient, with similar enough performance in most production tasks and gaming. You can certainly get into the weeds about the specific types of power efficiency one is better at than the other, but they are very comparable on the whole of power efficiency with a slight edge to Intel.
I’m not arguing anything about efficiency, just laughing at his comment… “stop the amd defense”
 
  • Like
Reactions: YSCCC
Mar 10, 2020
414
376
5,070
Looking at TPU and Gamers Nexus, the power use reduction 285 vs 14900k is admirable.

The many applications graph shows a large number of results with both the CPUs within spitting distance. There are also applications which show the 285 using appreciably less power than the 9950 and vice versa.

Gamers Nexus shows the 285 using less power than 14900k but its efficiency in the tests GN presents is not massively improved due to a drop in performance. The 9950 is more thirsty in some, more efficient in others. It does use more power but the time taken is less or in the compute tests GN used had greater throughput.

Work out what programs you want to primarily run, see which chip runs with the greatest throughput for those tasks. Look at the power for those tasks and then decide what is more important, energy used or time.
 
Last edited:
oh for the love of...

if tech power up and a number of reviewers got the power draw wrong. it's not their fault, i'm not saying they're bad sites.
Oh for the love of people who don't know what they're talking about at all. Congratulations you win the prize for writing a wall of text without knowing you're wrong, because rather than find out the truth you assumed you were right.

Asus
redesigned power delivery on some of their boards and this was noted in the TPU review:
The ASUS Z890 Hero motherboard feeds four of the CPU VRM phases from the 24-pin ATX connector, instead of the 8-pins, which makes it impossible to measure CPU-only power with dedicated test equipment (we're not trusting any CPU software power sensors). You either lose some power because only the two 8-pin CPU connectors are measured, or you end up including power for the GPU, chips, and storage when measuring the two 8-pin connectors along with the 24-pin ATX. For this reason, we used an MSI Z890 Carbon motherboard exclusively for the power consumption tests in this review.

edit: You also cited Roman, but clearly didn't actually watch his video on the subject. He even said that it might just be Asus because he didn't have any other motherboards in at the time.
 
Last edited: