ASRock X299 Taichi

[cadaveca]

yet I have to ask why show the DDR4-3600 if you did indeed test up to DDR4-4400?

The cost of such high-speed DIMMs is currently prohibitive to many, so I have a more "down-to-earth" speed that more people might be likely to buy. It's not about showing the best-case scenario, but rather relating what's realistic to what I hope is the majority of readers.

You can pretty much ignore that quote, I'm just pointing out that you've once again referred to testing overclocking.

Of course because, I mean, a lot of us want to push our CPUs a little bit too, right? So I found a power consumption level and clock that hopefully many can reach, too. But memory... memory clocking (I mean from top to bottom) has only small impacts on performance overall. To me, there are "gates" that offer incremental performance increases with memory, but those "gates" are tightly held by the cost of that memory.

I need to address several other points I saw now, so this post may be a little long. I'm addressing them in no particular order, either.

Great.

Just because elmor understands doesn't mean that everyone understands. Is it not good for someone who knows the right questions to ask said questions for those who don't? An average reader (I'm not patronizing the TPU membership, I'm talking truly average from whichever walk of life) may not know that the right questions even exist, so is it not a good idea for someone who knows the questions to ask them?

I'm also an overclocker, I'm also in elite league, so what? The review isn't aimed at us. We don't read a review and decide to buy a board. Reviews help those who are not in this position to choose a board, and chances are they don't have a fraction of the understanding to turn benchmark results into an informed decision. You're saying "these are the results, they're good" but not explaining the how or why.

That's a fair point for sure. A while back I discussed with W1zz going into more "basic" writing like you seem to be asking for, but guess what W1zz said? "We are an enthusiast site".

So, what defines an enthusiast? We will all come up with separate answers. However, when people do have questions, like you do, they can ask, and they can get answers, like you are. However, I know when someone's asking because they are generally interested, and when someone's asking just to be critical, so my responses adjust accordingly. You can see that pretty clearly in how I've responded here in this comment thread, even.

Then is the better option not to hold back the review until you can make a worthy comparison? Sure, you lose the "appeal" of being first with the review, but is a worthwhile review not worth more?

In some instances, yes. With this product, no. It's not like they are changing the VRM design or something... and it's not about being first, either. It was actually very interesting for me to write this review, have it not go live for quite some time, and then see some "issues" arise between the time that I kind of touched on in the review, like VRM cooling (although I hardly call think it's that big of a deal with some boards).

That is more than fair enough, but it doesn't explain how you can give it an Editor's Choice award when, for all anyone knows, it ends up being the worst X299 board out there. You've given this board a 9.8 out of 10 - honestly, what would a board score if it were to offer all of the features of the Taichi plus more (maybe another four SATA ports, easier overclocking, another M.2 slot, four extra USB 3.0 ports, whatever the case may be)? Would it score above 10, would the score of this board drop (as well as the Editor's Choice award fall away), or would it really only be able to score a MAXIMUM of 0.2 more than this? By giving it an Editor's Choice award straight off the bat you've limited your future review scoring unless you come back and edit this review, in which case was it reliable to begin with (rhetorical question)?

You mean, I drew a bar in the sand with my score? You bet I did. What crosses that line, and in which way, well, after so many years of doing this, what I expect should be pretty obvious. I mean, board makers send me specific boards for a reason; they know I should like them. It's not rocket science. That's why sometimes I cover products other sites don't, and OEMs send different boards to different sites.

Adding more features doesn't make a board better; let's make that clear. Also, a 9.8 for one board, doesn't mean it is equal to another 9.8 board... one might cost $200 more...

Then I have to ask what would set them apart? "MOAR RGB!!!"? There are only so many metrics that can be used to judge a board - appearance, I/O and performance being the main three. Appearance is too subjective, and I/O is something that can be read on paper (in which case list the specs of every X299 board and give the award to the board with the most I/O connectivity options). That leaves performance, which you want to leave out? Performance is one of the main things setting one board apart from another!

Yeah, you're right. But for me, performance comprises of two things; BIOS tuning and circuit design. When I start talking about that stuff, people don't really care to read it. We do know what pages people read...

I don't see what your point is. It's no secret at all who elmor is or where he works. The people who buy the boards he helps develop are not him, and 99.9 % don't have the understanding he has. Is it not in the best interests of everyone to make sure that the reviews are accurate and helpful to the 99.9 % ?

Sure, but this is a particular subject. He complained about the industry, it was covered here. He said boards sell fine though. I said maybe he's out of touch with his customer base. I mean, boards are selling...

But here's the old thinking, that performance is all there is, or has to be the most important factor, and unless you're competitive, that's not really 100% what the end user is after.

Pre-release testing has indeed got problems as we've already discussed - but pre-release is only handy to those who actually get the hardware pre-release. Once the release hardware is on the shelves and in Joe Soap's computer it isn't necessarily the same kettle of fish as the pre-release version, so is the review even applicable to Joe Soap? If you're aiming your review at Joe Soap you've got to make sure that your experience will as closely exactly match his as possible, which isn't possible with a pre-release review.

Again, we are an enthusiast site. I write for enthusiasts. They come in all shapes and sizes and colors and from all over the world, and they all want different things. I can't keep them all happy. But here, I can show early stuff, and then later, I can show how it's changed over time...

You're so very right, and that's why the dates are right in the beginning. So people know what's what here. But you seem to be expecting something else. I can't please everyone.

I think that's it for now.

I do value the feedback, even elmor's. I mean, we don't have to agree on everything; the world would be pretty boring. Ultimately, I do reviews to create discussions, no matter where they might go. Again, economy of words.

 

[[XC] Oj101]

The cost of such high-speed DIMMs is currently prohibitive to many, so I have a more "down-to-earth" speed that more people might be likely to buy. It's not about showing the best-case scenario, but rather relating what's realistic to what I hope is the majority of readers.

That would be a fair statement, if you didn't go on to say that this is an enthusiast site and the reviews are tailored accordingly. If it were indeed a Joe Soap review I'd say fine, maybe don't even test such a high speed.

That brings me to another point, though - the RAM benchmark results don't look as though they were all done with dual channel DDR4-3600 CL16 (in fact I'll say I'm positive they were not - the LGA1151 results are looking very much like DDR4-2133, while the LGA2011v3 results are certainly quad channel which X370 doesn't even support), so would you not say that the results of such are entirely misleading?

Are you testing the platform or the motherboard? If the platform, a single motherboard is not enough and DDR4-2133 on LGA1151 is not doing the older platform any justice when you're testing a memory speed two thirds faster (and not officially supported by the platform) on X299. If the motherboard, results need to be directly comparable and you should use settings which work across the board - something such as dual channel DDR4-2800 CL15 with a 100 Mhz strap where applicable and always with a 100 MHz base clock.

Of course because, I mean, a lot of us want to push our CPUs a little bit too, right? So I found a power consumption level and clock that hopefully many can reach, too. But memory... memory clocking (I mean from top to bottom) has only small impacts on performance overall. To me, there are "gates" that offer incremental performance increases with memory, but those "gates" are tightly held by the cost of that memory.

And I'm sure those same people like to push their CPUs a little too, is that not so? Memory clock speed can greatly influence overall performance even on a Skylake Core i3 - https://view.joomag.com/mag/0458172001466490420/p14

You don't have to take my word for it; others have also tested and seen silly big results such as 30+ % performance gains in Fallout 4 (that size performance increase is often not even possible between processor SKU jumps).

That's a fair point for sure. A while back I discussed with W1zz going into more "basic" writing like you seem to be asking for, but guess what W1zz said? "We are an enthusiast site".

You see, now this just contradicts what you've been saying about not overclocking memory, not testing faster RAM speeds, etc. The review comes across as halfway between aimed at Joe Soap and enthusiasts. Forgive me for the confusion, but I'm not sure who this is truly aimed at.

So, what defines an enthusiast? We will all come up with separate answers. However, when people do have questions, like you do, they can ask, and they can get answers, like you are. However, I know when someone's asking because they are generally interested, and when someone's asking just to be critical, so my responses adjust accordingly. You can see that pretty clearly in how I've responded here in this comment thread, even.

Are you saying that Jon is wrong to point out the flaws in the review methodology? In that case you need to say the same to me - I'm NOT reading motherboard reviews to choose my next motherboard, I haven't since around the turn of the millennium. My reasons for reading reviews are much the same as his, except I don't work directly at a vendor.

In some instances, yes. With this product, no. It's not like they are changing the VRM design or something... and it's not about being first, either. It was actually very interesting for me to write this review, have it not go live for quite some time, and then see some "issues" arise between the time that I kind of touched on in the review, like VRM cooling (although I hardly call think it's that big of a deal with some boards).

What about the BIOS issue you addressed on the previous page where you speculated that a BIOS update would fix it? If it wasn't speculation and you actually tested and saw it working at full speed, should it not have been mentioned in the review? If it has not yet been fixed, should you not have knocked points off? If it remains speculation should that not be mentioned in the review?

You mean, I drew a bar in the sand with my score? You bet I did. What crosses that line, and in which way, well, after so many years of doing this, what I expect should be pretty obvious. I mean, board makers send me specific boards for a reason; they know I should like them. It's not rocket science. That's why sometimes I cover products other sites don't, and OEMs send different boards to different sites.

Adding more features doesn't make a board better; let's make that clear. Also, a 9.8 for one board, doesn't mean it is equal to another 9.8 board... one might cost $200 more...

Are you saying that you know for sure that no motherboard, released or coming, can be better than this board by a margin of more than 0.2/10? That seems shortsighted to me.

You say that vendors send you products that they know you will like - does that mean the final score is subjective rather than objective?

I also added all the extra features assuming the price and all other variables remain the same - we're talking the same board with a ton of extra features at the same price point. What would that score if it existed? There are boards so far in the future that even the vendors don't know what the final review will look like - how can you set a score that leaves so little headroom for improvement with anything that comes in the future?

Yeah, you're right. But for me, performance comprises of two things; BIOS tuning and circuit design. When I start talking about that stuff, people don't really care to read it. We do know what pages people read...

You don't need to go into that level of detail, but at least touch base. You say the review is aimed at enthusiasts, but other than the lack of covering the basics I don't see it.

BIOS tuning is one thing that really sets one board apart from another - most of the other differences are superficial and comparable on paper.

Sure, but this is a particular subject. He complained about the industry, it was covered here. He said boards sell fine though. I said maybe he's out of touch with his customer base. I mean, boards are selling...

Boards selling isn't the same as customers ending up with what it right for them, or boards could even be selling better if customers were properly informed. I can see his point.

But here's the old thinking, that performance is all there is, or has to be the most important factor, and unless you're competitive, that's not really 100% what the end user is after.

Again, what else are you really reviewing? The rest can be seen on a spec sheet and literally compared on paper. It makes good filler content for a review, but there should be more to it that that.

Again, we are an enthusiast site. I write for enthusiasts. They come in all shapes and sizes and colors and from all over the world, and they all want different things. I can't keep them all happy. But here, I can show early stuff, and then later, I can show how it's changed over time...

Again you say enthusiast and again I just don't see it.

You do say, however, that you'll come back and show how things have changed. Does that mean that this board's score could, in theory, be dropped to something like 7.4 if far superior boards offering better value come out in the future?

You're so very right, and that's why the dates are right in the beginning. So people know what's what here. But you seem to be expecting something else. I can't please everyone.

Joe Soap probably doesn't know the exact release date, or that things can change in the weeks/days leading up to launch, so the date on its own isn't sufficient for him. A better idea would be a disclaimer saying that the review was performed on non-retail/pre-release hardware and might not be indicative of retail performance.

Again, economy of words.

Economy of words is one thing, leaving out critical details is another altogether though.

 

[ShockG]

All one can take away from this, is that there is clearly room for improvement.
Motherboard scores are generally high and generally overblown, because they don't mean anything along with the awards that go with them. It's simply "reviewers" farming out editorial with obviously inadequate testing data being shown/reported.
Boards are generally the same, what makes the biggest difference really when all things are working as they should (No S3 or S5 issues, reliable memory training at whatever voltage, clean power delivery, no current issues on USB ports or saturation of bandwidth etc.) is memory performance, support and tuning. Ultimately that's where the rubber meets the road especially for these lofty claims vendors make about RAID/driver performance and memory frequencies.
This is a difficult thing to get right and a lot of effort (by engineers) goes into this and it's rarely tested properly.

Arguing for simplified, less data driven reviews is not ideal as it doesn't provide any meaningful information to readers and their knowledge isn't growing whatsoever. Most reviews lack context and just thorough testing. Nobody is saying you need test everything, but whatever you do test, be thorough about it. Perhaps this is too much to ask for and what you do is a general overview which is fine. Then at the very least touch on these various aspects of a motherboard with a little more care. So if you have B-die 3,600 sticks and a board says it supports 4,400MHz. Then try it, that is "overclocking" which has a dedicated section or page already. Upping the CPU multiplier is not enough. That's the part you can leave out in fact as if the board can't do that it means it isn't working as that's CPU spec. There's zero information to be gained via a multiplier change and increase in core voltage. Either try Uncore/Mesh on INTEL or Infinity Fabric via DRAM on AMD, but actually do overclocking. Just an "oh btw Uncore reached 3200 at X voltage" with a CPU-Z screenshot or even better yet, a link to that actual frequency.

Same with discussing audio, if there were issues, what was the issue? Cross-talk, noise, interference when you move the mouse, from fans etc. No need for pop culture references. You can have them of course, but assume some people don't have your particular taste in entertainment.

Like you said, you've created conversation and that's important. I think we all expect better from you, especially on TPU because so many other sites are just lost causes. TPU motherboard and DRAM reviews can improve. Improve enough to match the other product reviews you have here, which are by and large reliable and have a bit more detail.

It's good you didn't' hide behind a ban hammer though like so many other major sites.
It's why I still visit TPU.

 

[cadaveca]

Are you saying that you know for sure that no motherboard, released or coming, can be better than this board by a margin of more than 0.2/10? That seems shortsighted to me.

In a way, yes, that is exactly what I am saying.

You say that vendors send you products that they know you will like - does that mean the final score is subjective rather than objective?

Not at all; it is simply the criteria that are used to judge products are pretty plain. The OEMs recognize it, how come you don't? Oh, that's right, you like do competitive benchmark, so an ideal board for you is something else...

I also added all the extra features assuming the price and all other variables remain the same - we're talking the same board with a ton of extra features at the same price point. What would that score if it existed? There are boards so far in the future that even the vendors don't know what the final review will look like - how can you set a score that leaves so little headroom for improvement with anything that comes in the future?

That depends... are the features superfluous? Do they match the overall theme of the board? I do see things as being "over the top" in some instances and I have no problem saying so.

BIOS tuning is one thing that really sets one board apart from another - most of the other differences are superficial and comparable on paper.

Most, yes. I mean, we do agree on a lot, and that's great.

Again you say enthusiast and again I just don't see it.

That's because we are different types of PC enthusiast. Not all enthusiasts want the exact same thing you do, or what I do. Like, me, I don't care for benchmarking at all, but you like it, what are we to do?

You do say, however, that you'll come back and show how things have changed. Does that mean that this board's score could, in theory, be dropped to something like 7.4 if far superior boards offering better value come out in the future?

To 7.4? Nope. The Taichi boards (and all boards, for that matter), are designed for a specific user. It's not a Fatal1ty board... What am I reviewing? LOL.

Joe Soap probably doesn't know the exact release date, or that things can change in the weeks/days leading up to launch, so the date on its own isn't sufficient for him. A better idea would be a disclaimer saying that the review was performed on non-retail/pre-release hardware and might not be indicative of retail performance.

I don't think you give our readers enough credit, and you don't visit here often enough. Our news section is ALL OVER launches and readers should know full well when something is coming. And again, I know far more about our readers and how our content is consumed than you seem to think...

Economy of words is one thing, leaving out critical details is another altogether though.

I don't see the things you see as critical in the same light. And gain, you like competitive benchmarking, so you looking for these things makes a lot of sense, but guess what happens when we make a news post about such content? Most responses will be to say how dumb it is, and a waste of time, and how they want things more "realisitic". You seem to forget that users still exists, and they don't want what you want.

So if you have B-die 3,600 sticks and a board says it supports 4,400MHz. Then try it, that is "overclocking" which has a dedicated section or page already. Upping the CPU multiplier is not enough. That's the part you can leave out in fact as if the board can't do that it means it isn't working as that's CPU spec. There's zero information to be gained via a multiplier change and increase in core voltage. Either try Uncore/Mesh on INTEL or Infinity Fabric via DRAM on AMD, but actually do overclocking. Just an "oh btw Uncore reached 3200 at X voltage" with a CPU-Z screenshot or even better yet, a link to that actual frequency.

I understand. However, let's compare my overclocking and W1zz,'s overclocking. All he does is adjust clock with no voltage changes. Technically, I'm doing more than he is, because I adjust clocks and voltages. So why is OK for W1zz, but not for me? Or perhaps, while you would like to see more, I've been told more is not needed?

Same with discussing audio, if there were issues, what was the issue? Cross-talk, noise, interference when you move the mouse, from fans etc. No need for pop culture references. You can have them of course, but assume some people don't have your particular taste in entertainment.

When there is something to discuss, I do, like with the ASUS MAXIMUS IX FORMULA...



Meh. I'll tell you what.. would you like to do reviews? If someone else will step up, I'll gladly go back to playing video games... That actually sounds like a great idea. Someone tell W1zz to post an ad on the front page.

 

[[XC] Oj101]

I saw this message a few days ago, but decided not to reply due to the extremely dismissive nature of your post. I've changed my mind, however, so here goes:

In a way, yes, that is exactly what I am saying.

OK, you must know something even the vendors don't - how can you be so sure of the upcoming boards that are still a way away, such as the Rampage or Apex?

Not at all; it is simply the criteria that are used to judge products are pretty plain. The OEMs recognize it, how come you don't? Oh, that's right, you like do competitive benchmark, so an ideal board for you is something else...

Absolutely - that doesn't mean I'll score an Apex or OC Formula board a perfect 10 or even anywhere near that - based on the lack of consumer features and high costs I'd probably score them around a 7 - lower than a Maximus Extreme, Formula or similar.

That depends... are the features superfluous? Do they match the overall theme of the board? I do see things as being "over the top" in some instances and I have no problem saying so.

I've listed potential features - eg more I/O connectivity. Would you score a board negatively if it had 12 SATA ports as that is "over the top" or would you praise it for offering something nothing else does?

Most, yes. I mean, we do agree on a lot, and that's great.

If you agree that the performance is actually an important factor in the review, why not go more in depth? Or at least make score comparable - the memory scores you have at the moment are just random numbers on a graph with no meaning to mos people. What memory speed and timings were used? What cache speed? Single, dual or quad channel? They're not mentioned, but I can clearly see that they aren't the same across the board. You're got speeds ranging from entry level DDR4 to mid/high end DDR4 in dual and quad channel all jumbled together. What favours does that do for the reader?

To 7.4? Nope. The Taichi boards (and all boards, for that matter), are designed for a specific user. It's not a Fatal1ty board... What am I reviewing? LOL.

7.4 was a randomly picked number. If a board comes out with far more features and offers double the value, how do you validate the score on paper when one scores 9.8/10 and the other 10/10? That appears to be a difference of a maximum possible of only 2 % between the Taichi and any other better board that comes out in the future - if a board is even just 10 % better how do you show that on paper when the score implies 2 % or less?

I don't think you give our readers enough credit, and you don't visit here often enough. Our news section is ALL OVER launches and readers should know full well when something is coming. And again, I know far more about our readers and how our content is consumed than you seem to think...

1. Your readership has nothing to do with the quality and thoroughness of a review. Whether you have one Joe Soap reader or 1,000 advanced readers, why should you skimp on parts of the review?
2. You keep changing your mind as to whether you're aiming at enthusiasts or Joe Soap - for example, you don't overclock the memory because that's too advanced, but you run it faster than the stock speed of the motherboard because that's too simple.
3. You have no idea how often I visit here. Would several times per day surprise you? Just because I'm rarely logged on doesn't mean I'm not here.
4. Considering there are simple driver related questions on the forum I think you're generalizing a bit too much as to how advanced your readers are.
5. You're not taking into account your non-member readership. I can promise you that you have far more readers than members, and you have no way of knowing anything about the knowledge of non-member readers.
6. "I don't visit here often enough" - does every other member? If not, you probably can't be very sure of their knowledge.

I don't see the things you see as critical in the same light. And gain, you like competitive benchmarking, so you looking for these things makes a lot of sense, but guess what happens when we make a news post about such content? Most responses will be to say how dumb it is, and a waste of time, and how they want things more "realisitic". You seem to forget that users still exists, and they don't want what you want.

Eh... Doesn't that confirm that the average reader is not an enthusiast? What is your definition of an enthusiast? What is an enthusiast who doesn't care about performance? How would that be different to a naive gamer? An enthusiast should care about these things whether he participates or not. It's the extreme guys who help bring once-extreme features to the mainstream.

A little more than a decade ago, in the Socket 939 days, Asus' top boards offered a maximum core voltage so low that you could run it and more 24/7 on air cooling, and if you wanted more you'd need to mod your board. Now we have voltages that are safe for 24/7 use only with chilled water, voltages that are only safe for a quick bench run, voltages only safe for validating a high frequency with CPU-Z, and even voltages that are relatively unsafe on LN2 - and these are available from relatively low end all the way up.

The extreme guys are the ones that make "auto" RAM settings work, as they have spent countless hours researching what does and doesn't work, as well as what is and isn't safe, in order for the BIOS to have auto settings which, for most people, will offer a better experience than they'll get by manually tweaking their system.

An enthusiast might not use auto, but should certainly appreciate the effort that goes into getting us where we are. A car enthusiast is pretty well defined and follows a similar trend in the car world, so why is the computer enthusiast any less? Is a computer enthusiast the equivalent of sticking an aftermarket wing on the boot, under-car lighting and a coffee can exhaust pipe, or is it something more?

If you want to aim at both Joe Soap AND enthusiasts, that's perfectly fine! BUT - then make it a usable review for both. As it stands it certainly isn't very usable to enthusiasts, and, to a large extent, to Joe Soap either.

I understand. However, let's compare my overclocking and W1zz,'s overclocking. All he does is adjust clock with no voltage changes. Technically, I'm doing more than he is, because I adjust clocks and voltages. So why is OK for W1zz, but not for me? Or perhaps, while you would like to see more, I've been told more is not needed?

I don't remember anyone saying that the method W1zz uses is perfect? You seem to be reading between lines which aren't there.

There was decent advice in the post you quoted, but you went off on a non-existent tangent :/

Meh. I'll tell you what.. would you like to do reviews? If someone else will step up, I'll gladly go back to playing video games... That actually sounds like a great idea. Someone tell W1zz to post an ad on the front page.

If you want to be a reviewer for the public you need to be able to take criticism as you are not perfect and not capable of making a perfect review - there has been a lot of constructive criticism but your approach to it is "screw it, I wanna be gaming anyway" If you took a second to actually listen to what's being said you might come away with the ability to write an absolutely kickass review