28c Intel W-3175X Reviews; $1700 motherboard?

[Tristerin]

10 minutes ago, leadeater said:

So what would stop people getting confused if the 2990WX OC results where also in the other graphs? What would stop them using OC performance and stock power?

 

Are you really going with the majority of the GN audience doesn't know how to read graphs nor knows what OC and stock is and are incapable of identifying it?

 

Well then you can find that out when the 2990WX review comes out, or find out using a different source who has a complete 2990WX review. Until then this is an Intel W-3175X review.

 

That has skewed data pareto'd off incorrectly.  The above statement you wrote, does not change that.

[Lathlaer]

@Tristerin the facts were never in dispute, only your interpretation of them.

 

Where are you getting the notion that they did it to cater to Intel?

[Tristerin]

Just now, Lathlaer said:

@Tristerin the facts were never in dispute, only your interpretation of them.

 

Where are you getting the notion that they did it to cater to Intel?

I gave two options - left off intentionally or accidently (you count my words, how many times did I say it could be a mistake sir).  Because the information is CLEARLY there if they have the Power Draw.

 

After that, its time for Tinfoil hats.  My tinfoil hat says most reviewers are human beings and are capable of favoring.  Are capable of mistakes also.  But when it comes to Business (Tech Reviewers are making money) and Capitalism...the amount of cloak and dagger is real and so my opinion is it was intentional.  No other reason to leave that data out.

 

So it comes down to this:

 

Graphs are not presented apples to apples

 

Some people are okay with that, I am not. 

[Fasauceome]

7 minutes ago, Tristerin said:

  I insinuated in my original post that the graphs are misleading

You actually stated that the graphs were misleading, overtly. You insinuated that GamersNexus was misleading, and that he's shill for Intel. Steve went after Principled Technologies back when they posted results that the i9 9900k was 50% faster than the R7 2700x, which had just wretched testing methodology. For you to claim Steve Burke is that kind of hypocrite made me insulted on his behalf, probably too strong a phrasing but if you're gonna make fun of me for being "triggered" then that's waaaay off topic, and childish derisiveness.

 

You can't pretend to have such a superior comprehension when you associate such things as 2 graphs = GN is bad. The original post includes what you feel to be misleading graphs, but since you didn't read or hear Steve's conclusion then why would you speak on the legitimacy of his review?

[leadeater]

10 minutes ago, Tristerin said:

Now we need to question why?  Is the why...because it was a mistake?  Is the why...because it torched Intels new chip? 

Maybe because someone thought people would like to know what the OC power draw is so included it but was not able to complete the required testing for the other roughly 12 tests so did not include any OC performance results, even if some of the data did exist. Is it any more acceptable to have only 1 test include the OC performance results but not the other 12?

 

Blender GN Logo just happens to be their standard CPU power test method, those two graphs aren't as directly related as you might be thinking. All the performance graphs for workstation applications are shown first then that power graph is shown.

 

Not only that the W-3175X OC power results are shown which is significantly more than the 2990WX OC power draw, 240W more which is more than the stock 2990WX power draw.

 

https://www.gamersnexus.net/hwreviews/3433-intel-xeon-w-3175x-28-core-cpu-review-benchmark-overclocking

 

[Tristerin]

4 minutes ago, fasauceome said:

You actually stated that the graphs were misleading, overtly. You insinuated that GamersNexus was misleading, and that he's shill for Intel. Steve went after Principled Technologies back when they posted results that the i9 9900k was 50% faster than the R7 2700x, which had just wretched testing methodology. For you to claim Steve Burke is that kind of hypocrite made me insulted on his behalf, probably too strong a phrasing but if you're gonna make fun of me for being "triggered" then that's waaaay off topic, and childish derisiveness.

 

You can't pretend to have such a superior comprehension when you associate such things as 2 graphs = GN is bad. The original post includes what you feel to be misleading graphs, but since you didn't read or hear Steve's conclusion then why would you speak on the legitimacy of his review?

It is misleading - insinuation using the modicum of communication that is text is the readers choice.  Never used the word shill.

 

Never claimed - so bro, I like you but you are going to need to back off.  Never used shill, hypocrite or insulting terms (stupid, dumb, idiots - I only talked about his Nuts) - and you are triggered by definition.  Read what you just typed to me man.

 

I don't have superior comprehension - my first post, was correct 100%.  Its misleading, as the graphs are not pareto'd correctly to show comparisons.  That still cant change because those are facts.  I have to question that, its who I am.

 

I didn't speak on his legitimacy of his reviews.  If I did please, post what I said about his legitimacy.  Or are you reading to far into what I type and getting mad at me for it?

 

4 minutes ago, leadeater said:

Maybe because someone thought people would like to know what the OC power draw is so included it but was not able to complete the required testing for the other roughly 12 tests so did not include any OC performance results, even if some of the data did exist. Is it any more acceptable to have only 1 test include the OC performance results but not the other 12?

 

Blender GN Logo just happens to be their standard CPU power test method, those two graphs aren't as directly related as you might be thinking. All the performance graphs for workstation applications are shown first then that power graph is shown.

 

Not only that the W-3175X OC power results are shown which is significantly more than the 2990WX OC power draw, 240W more which is more than the stock 2990WX power draw.

 

https://www.gamersnexus.net/hwreviews/3433-intel-xeon-w-3175x-28-core-cpu-review-benchmark-overclocking

 

This. Right. Here.  We can discuss this -

 

For you first paragraph - I cannot come to a logical conclusion of why these results were not included.  They should be.  Someone mentioned earlier in this thread something about a barrowed TR - could he NOT OC it?  Because he has the power draw results...which means he either did...or took someone (or some previous) tests information.  This should be heavily denoted imho.

 

I agree that the power is significant over TR...but what can the TR do with the power results he has...why wasn't that posted?  Seems cloak and dagger to me.  Probably because its Capitalism and I don't trust people inherently when money is involved because $$ makes people do stupid stuff.

 

Would really love to see how the OC TR competes in like environments against this new insanely expensive intel chip.

 

 

[Fasauceome]

10 minutes ago, Tristerin said:

If I did please, post what I said about his legitimacy

These ring a bell?

10 minutes ago, Tristerin said:

Its why I don't watch certain reviewers.  Its misleading by default.

 

23 minutes ago, Tristerin said:

But when it comes to Business (Tech Reviewers are making money) and Capitalism...the amount of cloak and dagger is real

 

10 minutes ago, Tristerin said:

They still want their samples from Intel so they are playing the "game".

GN gets their samples without playing any game, people want GN reviews and when they don't get them it's a problem. Nvidia didn't send a review sample of the 2060 to Steve, and people had to raise some hell about it. He actually bought one himself to do the testing, he's not out there pimping these these companies' products but he's being honest. He literally makes "Dissapointment PC" videos where he takes the biggest let downs of the year and builds a PC out of them, talking trash on such things as the i9 9th gen HEDT and Vega FE, and guess what? He still gets samples from Intel and AMD for his reviews, because he's a fair and honest reviewer.

 

So yes, you did call into question the quality of GamersNexus, and yes, you did so erroneously. As for Steve saying he does have to retest the 2990WX with a stable OC for more data, that's even better. I still feel as though reading the actual GN review or watching it gives a far more complete picture, again stating that the OC power draw was indeed an afterthough just to have some more data (and it's not a damn pop up book, you gotta read some words too).

 

Additionally, the original post of this topic, which is your main point of contention, reaches a pretty neutral conclusion along the lines of "wow it's fast huh?" so having a small gripe with that would probably be fine but you really went out of your way to target the source of the graphs.

[Lathlaer]

5 minutes ago, Tristerin said:

I agree that the power is significant over TR...but what can the TR do with the power results he has...why wasn't that posted?

I have to ask - you really expect some crazy miracle there? After seeing how stock Intel in Blender did vs. OC'ed Intel in Blender, you expecting some big extraordinary revelation?

 

I mean, when I glanced over the specifications of both chips (base clock speed vs. turbo vs. OC) and saw the difference from Intel's 3.1GHz base to 4.5GHz in Blender, I don't really expect TR to suddenly be 500% better or smth.

 

Truth is, considering all the facts about Blender differences stock vs. OC of both chips and the huge difference of power draw, I just don't see it.

 

The likely scenario will be that which can be reasonably expected from what we have - that OC'ed TR will do a bit better than stock TR, maybe a bit better than stock Intel but worse than OC'ed Intel. It hardly changes anything because if he really wanted to fudge results to favor Intel, Blender is a terrible place to do it.

[Tristerin]

Just now, Lathlaer said:

I have to ask - you really expect some crazy miracle there? After seeing how stock Intel in Blender did vs. OC'ed Intel in Blender, you expecting some big extraordinary revelation?

 

I mean, when I glanced over the specifications of both chips (base clock speed vs. turbo vs. OC) and saw the difference from Intel's 3.1GHz base to 4.5GHz in Blender, I don't really expect TR to suddenly be 500% better or smth.

 

Truth is, considering all the facts about Blender differences stock vs. OC of both chips and the huge difference of power draw, I just don't see it.

 

The likely scenario will be that which can be reasonably expected from what we have - that OC'ed TR will do a bit better than stock TR, maybe a bit better than stock Intel but worse than OC'ed Intel. It hardly changes anything because if he really wanted to fudge results to favor Intel, Blender is a terrible place to do it.

 

No I expect no crazy miracle, I espouse that these tech tubers we heavily rely on give us the best information they can.  Because we support them which puts food on their tables.  Its not that I expect something extraordinary, I just want apples to apples comparison that doesn't have the appearance of leaving out information

 

The information would then (which it is relevant, the OC potential in Blender for the TR) allow US to further extrapolate out the delta in the Power to Performance ratio (OC = free performance...though not guaranteed blah blah blah) vs Costs.  I think these are very relevant when we are talking about a $3000 chip.

 

Your last paragraph - "Likely scenario" - I mean they have the hardware...the platform...the audience...I questioned why it was left out.  That's all I did, and here we are lol.  That delta is in reality then explainable further into $$$ per millisecond saved - which is what true Workstation users (power users) need/want, no?

[leadeater]

22 minutes ago, Tristerin said:

For you first paragraph - I cannot come to a logical conclusion of why these results were not included.  They should be.  Someone mentioned earlier in this thread something about a barrowed TR - could he NOT OC it?  Because he has the power draw results...which means he either did...or took someone (or some previous) tests information.  This should be heavily denoted imho.

I mentioned it, he had to return it as stated in the review. The data does not exist, it's incomplete but there is power draw data for the 2990WX that is usable so was included.

 

Almost all the other power draw data is from previous data, you actually think he retests all those CPUs every time? In fact all the data that isn't the 3175X and 2990WX are previous test data, with the exception of the Adobe Premiere test where new sample footage was used and only 4 CPUs were tested.

 

22 minutes ago, Tristerin said:

This should be heavily denoted imho.

It is, in the written review and in the video.

 

I don't know about where you work but graphs are not present alone without a presenter to explain them or not contained in a full report. All you're doing is taking images/information out of context and interpreting however you wish to, which is itself misleading. Not that people don't do this all the time, cherry picking graphs to support an argument irrespective of what the graph was actually for or how it was being used. 

 

22 minutes ago, Tristerin said:

why wasn't that posted?  Seems cloak and dagger to me.  Probably because its Capitalism and I don't trust people inherently when money is involved because $$ makes people do stupid stuff.

Because not all the testing is complete and including OC results in only some of the graphs is unacceptable, there is zero need to rush results for a product not being tested/reviewed at the time. When the 2990WX review comes out it will also have all the W-3175X data.

 

22 minutes ago, Tristerin said:

Would really love to see how the OC TR competes in like environments against this new insanely expensive intel chip.

Then wait for the 2990WX review and you'll get it. Spoiler, look at TR1. OC won't do much at all. It doesn't for TR1/TR2 and it doesn't for Ryzen 2000 either.

[Tristerin]

28 minutes ago, fasauceome said:

You actually stated that the graphs were misleading, overtly. You insinuated that GamersNexus was misleading, and that he's shill for Intel. Steve went after Principled Technologies back when they posted results that the i9 9900k was 50% faster than the R7 2700x, which had just wretched testing methodology. For you to claim Steve Burke is that kind of hypocrite made me insulted on his behalf, probably too strong a phrasing but if you're gonna make fun of me for being "triggered" then that's waaaay off topic, and childish derisiveness.

 

You can't pretend to have such a superior comprehension when you associate such things as 2 graphs = GN is bad. The original post includes what you feel to be misleading graphs, but since you didn't read or hear Steve's conclusion then why would you speak on the legitimacy of his review?

 

7 minutes ago, fasauceome said:

These ring a bell?

 

 

GN gets their samples without playing any game, people want GN reviews and when they don't get them it's a problem. Nvidia didn't send a review sample of the 2060 to Steve, and people had to raise some hell about it. He actually bought one himself to do the testing, he's not out there pimping these these companies' products but he's being honest. He literally makes "Dissapointment PC" videos where he takes the biggest let downs of the year and builds a PC out of them, talking trash on such things as the i9 9th gen HEDT and Vega FE, and guess what? He still gets samples from Intel and AMD for his reviews, because he's a fair and honest reviewer.

 

So yes, you did call into question the quality of GamersNexus, and yes, you did so erroneously. As for Steve saying he does have to retest the 2990WX with a stable OC for more data, that's even better. I still feel as though reading the actual GN review or watching it gives a far more complete picture, again stating that the OC power draw was indeed an afterthough just to have some more data (and it's not a damn pop up book, you gotta read some words too).

 

Additionally, the original post of this topic, which is your main point of contention, reaches a pretty neutral conclusion along the lines of "wow it's fast huh?" so having a small gripe with that would probably be fine but you really went out of your way to target the source of the graphs.

 

These Ring a Bell:

 

      17 minutes ago, Tristerin said:

Its why I don't watch certain reviewers.  Its misleading by default.

 

31 minutes ago, Tristerin said:

But when it comes to Business (Tech Reviewers are making money) and Capitalism...the amount of cloak and dagger is real

 

18 minutes ago, Tristerin said:

They still want their samples from Intel so they are playing the "game".

 

Given the left out information provided by GN, what above that you took the time to peruse was so insulting?  Again, this isn't insulting, this is discussing why it was left out.  You have a heavy like for this guy, so it must sound insulting.  I don't know, I don't see the insult.  Maybe I come from a time when an insult meant you should punch that jerk in the mouth.  If these are insults, than lots of fists would be flying.

 

Odd how I only said They once and that would be the only comment you could insinuate (bridge the gap of actual facts typed out) was talking about GN in a negative light.  What I said wasn't even close to insulting. 

 

4 minutes ago, leadeater said:

I mentioned it, he had to return it as stated in the review. The data does not exist, it's incomplete but there is power draw data for the 2990WX that is usable so was included.

 

Almost all the other power draw data is from previous data, you actually think he retests all those CPUs every time? In fact all the data that isn't the 3175X and 2990WX are previous test data, with the exception of the Adobe Premiere test where new sample footage was used and only 4 CPUs were tested.

 

It is, in the written review and in the video.

 

I don't know about where you work but graphs are not present alone without a presenter to explain them or not contained in a full report. All you're doing is taking images/information out of context and interpreting however you wish to, which is itself misleading. Not that people don't do this all the time, cherry picking graphs to support an argument irrespective of what the graph was actually for or how it was being used. 

 

Because not all the testing is complete and including OC results in only some of the graphs is unacceptable, there is zero need to rush results for a product not being tested/reviewed at the time. When the 2990WX review comes out it will also have all the W-3175X data.

 

Then wait for the 2990WX review and you'll get it. Spoiler, look at TR1. OC won;t do much at all. It didn't for TR2 and it doesn't for Ryzen 2000 either.

 

(first two paragraphs) - as long as the testing environment is the same that's completely acceptable.

 

(on being heavily denoted) - its a shame they didn't do it on the graphs, would have made this a VERY well done review.  This caused me to question the entire things legitimacy, and in the business world is a huge no no.  Perhaps I shouldn't hold these guys to the candle Im held to in the business world.

 

I work for Big Daddy Warbucks.  (Warren Buffet).  You aren't allowed these types of mistakes.  Everything is HEAVILY vetted prior to presentation.  You aren't allowed these mistakes - again probably holding the GN guys to the candle I shouldn't be perhaps.

 

IMHO - this information (the OC power draw) should have been left out, I do get that (my opinion)

 

That's all I questioned that started this lol.  I think it was a bad move (again, given the environments I thrive in, I am thusly a product of it as described above)

[leadeater]

24 minutes ago, Tristerin said:

(on being heavily denoted) - its a shame they didn't do it on the graphs, would have made this a VERY well done review

To what, make the graph more messy? Or to put unreadable text on it? Clear concise graphs are far and away better and less susceptible to being misread or ignored.

 

There is nothing that could be added to the graph that could make it better. To put on what you want that would be adequate would be far to much text, of which is acceptable to put before or after the graph that explains what it is.

 

You don't write a thesis on a graph, graphs display data not explain them. They are supporting evidence.

 

Every graph even directs you to actually read the review which contains the required information, it's right there on all the graphs.

 

24 minutes ago, Tristerin said:

IMHO - this information (the OC power draw) should have been left out, I do get that (my opinion)

Well I did actually ask you that but I don't recall you ever answering it or saying this. However you have not raised this point before in this discussion, instead opting for accusing of misleading and inaccurate information.

 

My opinion is that it would make zero difference in or out, it doesn't make any of the other data misleading.

24 minutes ago, Tristerin said:

That's all I questioned that started this lol.  I think it was a bad move (again, given the environments I thrive in, I am thusly a product of it as described above)

But you didn't question, you accused and in any environment that tends to be unacceptable when done so without any evidence to support it and have not read the source information you are basing the accusation on.

 

All that aside the only point I have tried to clarify is on that including OC power draw is misleading and I have yet to hear a compelling argument to support that.

[Tristerin]

35 minutes ago, leadeater said:

To what, make the graph more messy? Or to put unreadable text on it? Clear concise graphs are far and away better and less susceptible to being misread or ignored.

 

There is nothing that could be added to the graph that could make it better. To put on what you want that would be adequate would be far to much text, of which is acceptable to put before or after the graph that explains what it is.

 

You don't write a thesis on a graph, graphs display data not explain them. They are supporting evidence.

 

Every graph even directs you to actually read the review which contains the required information, it's right there on all the graphs.

 

Well I did actually ask you that but I don't recall you ever answering it or saying this. However you have not raised this point before in this discussion, instead opting for accusing of misleading and inaccurate information.

 

My opinion is that it would make zero difference in or out, it doesn't make any of the other data misleading.

But you didn't question, you accused and in any environment that tends to be unacceptable when done so without any evidence to support it and have not read the source information you are basing the accusation on.

 

All that aside the only point I have tried to clarify is on that including OC power draw is misleading and I have yet to hear a compelling argument to support that.

I have to take the screenshot as a great denotion, however Asterisked denotions are very much a norm.  Cant argue with that SS though. 

 

However, imho, that doesn't make it clear by adding the OC power results of the TR...makes me question the reasoning.

 

I still stand by the data being misleading.  You don't include results that aren't tested, that can be tested, if you are only going to mislead with partial information.  Leave out the partial then.

 

My opinion is it makes a huge difference by further allowing us to see how BS this intel chip is.

 

"You accused" - this is my opening statement verbatim:

 

Why wasn't the OC'd Threadripper tested in the first graph...but its power consumption smack in the middle as an Overclock for power draw - looks like intentional placement to try to make it look like the Threadripper used a ton of power as well, to reach the score in the first picture.

 

Its why I don't watch certain reviewers.  Its misleading by default.

 

When in reality the Threadripper smashes.  In the test and power draw and in the price bracket.  Still.

 

This is misleading by leaving our information.  I find it misleading.  I noticed it the moment my eyes laid on the graphs that it was missing information.  No reason to leave this info out of the graphs, if you are going to put some of the info on some of the graphs. 

 

The stance I have taken, across tons of people accusing me of all kinds of things, is that the graphs are incorrect, misleading, and should have been done better.

 

THAT IS STILL THE CASE. (my stance)

 

However, I do appreciate the screenshot because it forgives a portion of it that you CAN find the information of WHY the put in skewed graphs. 

 

Now for the meat and tatters of our (yours and Is only) discussion:

All that aside the only point I have tried to clarify is on that including OC power draw is misleading and I have yet to hear a compelling argument to support that.

 

When you peruse the graphs, quickly, for someone like myself (however I noticed immediately) who doesn't have time to dig into the meat and potatoes of a review, I can get the entire information based on Facts presented in the graphs (if in fact the graphs are showing all the Facts - in this case...its not its putting out information that is either irrelevant because we cannot begin to understand what that power draw from the TR in fact produced in the Blender test - or left out intentionally, et a la TIN FOIL HAT TIME!)

 

Yes it is the readers fault for moving past this and not really investigating the graphs...they should have left this info out (OC TR Power Draw).  Its inconclusive and convoluting the tests.  Doesn't help anything by showing an OC'd TR produces this much power vs an OC'd new Intel chip producing this much power for Intels Blender results only.  Give us AMDs results or not at all. 

 

My 2 copper coins.

 

 

*or are you telling me that there is no info on a Blender Test for an OC'd TR chip?  That it cant do it at all?* because that would be the only acceptable reason to leave this information out when putting in half of the information!

[Hellion]

Is this a joke?

 

You can literally buy 2x 2990WX chips + 2x X399 motherboards for the same price as one of these + the corresponding board and pretty much every benchmark it's a single digit percentage difference either way.

 

Intel continues full speed ahead on the fail train.

[Ashleyyyy]

i love how there is a massive argument on page 4 and 5, i click on this one and it's all peaceful again  

[Taf the Ghost]

27 minutes ago, firelighter487 said:

i love how there is a massive argument on page 4 and 5, i click on this one and it's all peaceful again  

I should probably have turned off thread notifications before they got going. I got a LOT.