Quick News


I am very excited about this. Shame that by then I will probably have had my fill of Warhammer II for a while.

My current Mortal Empires Campaign only as two races left. The Dwarves are taking the world by storm and own almost 100% of the map. The Orcs, Undead, Chaos, Bretonians, Norsemen, Empire, Wood Elves and Dark Elves have already been wiped out. Only the Lizardmen and High Elves (Almost gone) remain.
 
Meltdown is the Intel one that will bugger it. Spectre is the one that affects all, but apparently has hardly any performance impact.
 
Why Raspberry Pi is not vulnerable to spectre and meltdown, along with nicest explanation of the vulnerability so far:

https://www.raspberrypi.org/blog/why-raspberry-pi-isnt-vulnerable-to-spectre-or-meltdown/
 
https://www.youtube.com/watch?v=CSI6N6RKd5A&t=131s

And so it continues. So not only does CL boost to speeds it shouldn't be (IE what is said on paper) but it also throttles to a different TDP after 15 seconds. So basically for 15 seconds it clocks to 4.3ghz, then drops to 3.7 to meet the TDP.

"Pulling about every trick you can" comes to mind.

ekCKNBo.png


Taken directly from Intel's own info.

zaf8LnD.png


Or to those that don't understand this. What this basically means is that the 8700 for example will run at different power levels (or clocks) depending on where it resides. So if it is in a HEDT board like a Asus for example you can probably disable it and the TDP. However, in other rigs like that one in the video the OEM can decide where to set it. So you could see a 1ghz discrepancy. So basically you go to Linus, watch a "review"of the 8700, decide it's the CPU for your new rig then you buy a pre built and... Absolutely nothing like it was when "advertised". Noting that I have replaced the word "reviewed" with "advertised" because we all know that Coffee Lake was boosting beyond spec in the initial "reviews" we got of it and thus is not as good as people think it is.
 
Last edited:
So is this a Coffee Lake Con or a motherboard/system manufacturer con? Intel does list the specs clearly on their website, as in single core Turbo frequency and baseline frequency.

Multicore Enhancement isn't Intel technology. It's an automatic overclock, and along with it TDP goes out of the window.

Reviewers are at fault if they report MCE results - and good reviewers make it clear what MCE does, like TTL.

Here's the competition "pulling every trick they can":
cbFpcoc.jpg
 
Last edited:
Yes, the specs are clearly listed. However, they are not clearly explained by any one.

The 8700 has a base clock of 2.8ghz. Depending where it is and what you are doing with it it can boost as high as 4.8ghz for X seconds (10, IIRC) or even as little as a few milliseconds.

So, depending on where your 8700 may be (despite the reviews you read (bent) ) you could see a 2ghz difference, depending on the OEM, board and etc.

So basically all of these "reviews" that we see are all BS. What they really are are glorified advertisements, in the best possible scenario (see also : Coffee Lake Conjob video 1) and we are being screwed.

As for AMD? can you please stop that? seriously? Allow me to explain why.

All you are doing is making yourself look like a deluded fanboy/girl without addressing the actual issues. This is the third time some one has thrown me an AMD shaped curve ball in the past two days when I am complaining directly about something with Intel, not AMD. And the more you keep throwing these curve balls the more you distract from the actual problem at hand.

Honestly, it's almost like Intel have their own bots and droids out there making all of their bad behaviour vanish.

So instead of just jumping in feet first with "AMD TOO !" please discuss the matter at hand.

If you think this problem is anywhere even remotely near as bad as the CL problem? then discuss it, but not as a defence strategy.

The fact is that people who game on PCs are going to get ripped off, and all you can do is drag AMD into it too.
 
Poor thermal and VRM design by OEMs is absolutely nothing new - throttling is commonplace in budget market PCs. The worst accusation here is making room for inadequate cooling in spec.

And yeah I do drag AMD here, though their largest turbo variance seems to be 3.0 to 3.7, which is nowhere near as drastic. And all the reviewers use adequate motherboards, airflow and coolers, so yeah of course it also reaches the 3.7 despite staying at 3.0 or even throttling when slammed to the cheapest OEM motherboard with a lowest end cooler they could get away with.
 
Intel's boost clocks are far stronger and far more dependent on factors unknown to many casual PC users. The 1800X boosts to a maximum of 4.1Ghz on a single core. That's the CPU's max rated speed for a single core. It is dependent on many external factors of which AMD have been quite upfront about. They discussed at length what XFR was and how it would work, and the tech press have shown it to be not that beneficial for a lot of users. Most reviewers would suggest an R7 1700 or R5 1600 over the X equivalent and simply applying your own overclocks, which even the stock coolers are capable of handling.

The i7-8700 on the other hand has relied on reviewers who have allowed the CPU to boost significantly higher than the TDP allows and not specified it. The same applies to the i5-8400. I personally don't see it as being as big of an issue as others as the processors even under less than ideal circumstances (OEM builds with poor cooling) still perform well in most situations, but the point remains: Coffee Lake has been shown to have a somewhat unrealistically commanding position in gaming and other single-threaded workloads, to the point of some reviewers suggesting the 'K' SKUs should be "ignored".

The i5-8400 is rated to boost to 4Ghz on one core and 3.8Ghz on all four cores when under the right circumstances, but it can boost higher if the motherboard tells it to. The same applies to the 8700 which can overclock itself with the right motherboard to a whopping 4.6Ghz or more. This is way above Ryzen, but the base clocks are actually lower in some cases. That's a huge increase. Compare that to the R5 1600 which has a base clock of 3.2Ghz and a max boost of 3.6Ghz irrelevant of the motherboard and even with the stock cooler (as long as it's the stock one or above).

To me, this is a reviewer's issue. Intel's boosting works similar to Nvidia's GPU Boost 3.0. When the tech press review a GPU, they always (or should if they are worth the bandwidth they use) display the clock speed over time. When a reviewer doesn't list the 8700 boosting to 4.6Ghz in a game that doesn't care about core/thread count, and that clock speed cannot be achieved by the majority of users, that's not accurate representation. Just like I don't think a lot of graphics card reviewers show accurate results these days, including TTL. They refuse to update their graphs with modern drivers taken into account. That is misrepresenting actual performance.
 
Poor thermal and VRM design by OEMs is absolutely nothing new - throttling is commonplace in budget market PCs. The worst accusation here is making room for inadequate cooling in spec.

And yeah I do drag AMD here, though their largest turbo variance seems to be 3.0 to 3.7, which is nowhere near as drastic. And all the reviewers use adequate motherboards, airflow and coolers, so yeah of course it also reaches the 3.7 despite staying at 3.0 or even throttling when slammed to the cheapest OEM motherboard with a lowest end cooler they could get away with.

My issue is not with the specs. I think you miss my point. My issue is that none of the spec is being openly discussed and pointed out by the "reviewing" industry (AKA advertising).

If Linus, Jay and etc (not Tom, not bringing him into this) all explained it and pointed it out we'd be much better off. Fact is you would be better off with a Ryzen 1700 based rig, as it would not drop below 3ghz and therefore if you are going bone stock it would be the much better choice.

But we're not, are we? no. Instead we are not being told this and to that ends some poor sod is going to end up with a rig that he thought could do 140 FPS in Crysis 3 only to find it does about half of that.

So the problem is not the product, it's the false advertising "best case scenario but we are not going to explain that to you" part of the product. I explained this elsewhere about my experience with an EK Coolstream radiator. It was bloody awful. Yet no one had actually said that.

Intel's boost clocks are far stronger and far more dependent on factors unknown to many casual PC users. The 1800X boosts to a maximum of 4.1Ghz on a single core. That's the CPU's max rated speed for a single core. It is dependent on many external factors of which AMD have been quite upfront about. They discussed at length what XFR was and how it would work, and the tech press have shown it to be not that beneficial for a lot of users. Most reviewers would suggest an R7 1700 or R5 1600 over the X equivalent and simply applying your own overclocks, which even the stock coolers are capable of handling.

The i7-8700 on the other hand has relied on reviewers who have allowed the CPU to boost significantly higher than the TDP allows and not specified it. The same applies to the i5-8400. I personally don't see it as being as big of an issue as others as the processors even under less than ideal circumstances (OEM builds with poor cooling) still perform well in most situations, but the point remains: Coffee Lake has been shown to have a somewhat unrealistically commanding position in gaming and other single-threaded workloads, to the point of some reviewers suggesting the 'K' SKUs should be "ignored".

The i5-8400 is rated to boost to 4Ghz on one core and 3.8Ghz on all four cores when under the right circumstances, but it can boost higher if the motherboard tells it to. The same applies to the 8700 which can overclock itself with the right motherboard to a whopping 4.6Ghz or more. This is way above Ryzen, but the base clocks are actually lower in some cases. That's a huge increase. Compare that to the R5 1600 which has a base clock of 3.2Ghz and a max boost of 3.6Ghz irrelevant of the motherboard and even with the stock cooler (as long as it's the stock one or above).

To me, this is a reviewer's issue. Intel's boosting works similar to Nvidia's GPU Boost 3.0. When the tech press review a GPU, they always (or should if they are worth the bandwidth they use) display the clock speed over time. When a reviewer doesn't list the 8700 boosting to 4.6Ghz in a game that doesn't care about core/thread count, and that clock speed cannot be achieved by the majority of users, that's not accurate representation. Just like I don't think a lot of graphics card reviewers show accurate results these days, including TTL. They refuse to update their graphs with modern drivers taken into account. That is misrepresenting actual performance.

Spot on fella. AMD were more than revealing and up front with their processors and how they worked. In fact, they even made videos showing you how they worked, what speeds to expect and how to OC using Wattman.

So we all knew from day one what to expect from Ryzen. As you say, Tom explained it all perfectly. Only because AMD had given him that info, and not hidden it in the specs and small print.

My issue is not that a 8700 can sometimes in worst case scenarios only run at 2.8ghz depending on the silicon, your board, your PSU and so on. My issue is that we are not being told and having it all explained to us that sometimes the 8700 can end up being far worse than a stock 1700.

You know, at one point a few weeks ago my mate was looking for a rig for his cousin. Ironically the two rigs that ended up on the short list from a really bloody long one were both HP Omens. One had a 1700, the other an 8700. Now I can't remember for the life of me why I decided to push him to get the AMD. I kinda knew it wasn't going to be as good at gaming (it did have a slightly better GPU, though) but hey, all's well that ends well I guess.

Because that Ryzen rig is guaranteed to run 300mhz or so faster even in the worst case scenario.

Intel are just making it all incredibly complex, and giving themselves the means to shovel off shoddy silicon. I bet most of it will end up in the hands of OEMs like HP, too, who will get OEM trays of the things.

The same trays that 8 Pack basically lives on. He will take each and every single one of those CPUs and then mark it and categorise it into what it can do. And then we will pay more for it.
 
Last edited:
Back
Top