Slashdot Mirror

The intention was never to estimate the cost of Apples part.

My original claim was ".. and for gaming a Ryzen 3 2200G would work better and cost $100"

The point was never to tell what an i3 8100 cost or what the Apple system should cost or what building one yourself would cost.
The point was to say that I thought it was a stupid choice of CPU (specifically said for gamers) and that one could get a better system for less.

The reason the i3 8100 is a bad choice for gamers is that the Vega 8 or Vega 10 integrated graphics in the 2200G and 2400G have a higher performance. The 2200G would have a slightly lower performance as CPU as the i3 8100 but for gaming on integrated graphics that wouldn't be the problem but the problem would be the weak graphics performance of the i3 8100. The i3 8100 could had been a better choice for games if one used a discrete graphics card together with it but I doubt that's the case for the Mac Mini.
The 2400G adds SMT as in 2 threads per core and a bit larger GPU but since it cost 60% more it may not be worth it. Up to each and every one to decide but you'll getting into a territory where something like the i3 8100 + GT 1030 could be bought instead where the graphics part would be just about the same though you still have to spend more money on the i3 8100 + GT 1030, add even more money and get the 1050Ti instead and you're above the performance of those two Ryzen APUs but I doubt the Mac Mini have that.

As a cheap compact system my choice would definitely have been the 2200G or possibly 2400G instead, maybe both as options for the customer to decide. The i3 8100 is a stupid choice.

Sweclockers, https://www.sweclockers.com/te...
BF1 720p low, average fps:
2400G: 83
2200G: 77 (150% faster)
8100: 31
BF1 720p low, 1% low FPS:
2400G: 66
2200G 55 (110% faster)
8100: 26

CS:GO 720p high, average FPS:
2400G: 148
2200G: 137 (110% faster)
8100: 64
CS:GO 720p high, 1% low FPS:
2400G: 105
2200G: 105 (210% faster)
8100: 34

Dota 2 720p high, avg:
2400G: 84
2200G: 83 (130% faster)
8100: 36
lows:
2400G: 46
2200G: 46 (100% faster)
8100: 23

Fallout 4 720p low, avg:
2400G: 56
2200G: 52 (130% faster)
8100: 23
lows:
2400G: 48
2200G: 44 (130% faster)
8100: 19

Fortnite: Battle royale 720p low, avg:
2400G: 124
2200G: 111 (180% faster)
8100: 40
lows:
2400G: 93
2200G: 82 (170% faster)
8100: 30

Sweclockers also include 1080p results but that likely hurt the 8100 even further if anything since it's the weaker one. So the 2200G which is cheaper than the i3 8100 perform 100-180% better in games or 2-3x as good.

Cinebench multi-core:
2200G just below i3 8100 in performance, 2400G more than 1/3 faster.
x264 encoding:
2200G slower than i3 8100, 2400G faster.
7-zip compression:
2200G a tiny bit slower than i3 8100, 2400G faster.
Blender:
2200G a bit slower than i3 8100, 2400G faster by 30% of the time.
Photoshop Lightroom Export:
i3 8100 a bit faster than both the 2200G and 2400G
Veracrypt:
Both 2200G and 2400G beat the i3 8100, the 2400G beat it by 70%.

So as can be seen for a small system with integrated graphics there's not even a contest; the AMD chips are way better. For gaming so much that the i3 8100 doesn't belong in any system whatsoever. For non-gaming tasks the 2400G still beat the i3 8100 by a decent margin so that chip should had been avoided.

No, they unlikely aren't.

I know a bunch of people speculate completely out of their asses that that would be the reason Anandtech results are different to most others but that's likely the case because any serious tester would use the latest Windows update, BIOS and drivers.

You just claim that out of nothing. You shouldn't do that. You don't know that's the case. Just a bunch of idiots claim that like if it was the truth.

https://www.sweclockers.com/te...
720p medium GTX 1080Ti:
BF1: 8700K 25% better avg. 19% better 1% low.
Witcher 3: 8700K 21% better avg. 14% better 1% low.
Fallout 4: 8700K 18% better avg. 17% better 1% low.
Total War: Warhammer II: 8700K 17% better avg. 19% better 1% low.
Farcry 5: 8700K: 17% better avg. 18% better 1% low.

Anandtech is investigating why their results are different. Their 8700K results for GTA V seem to be higher than their earlier ones so it don't seem to be an issue with this specific run and maybe ASUS Crosshair VII Hero with G.Skill SniperX 2933 MHz with the AMD Ryzen 7 2700X simply run amazingly well. Or their results using 8700K is poor both now and before because they use only Crucial Ballistix 2666 MHz RAM and a Silverstone AR10-115XS cooler for the i7 8700K. Or some firmware or driver on the ASRock Z370 Gaming i7 board make it run bad.
I haven't compared the Anandtech results with similar machines with the same games and settings from some other reviewer so I can't tell if the Anandtech results are different from theirs.

If the memory kit AnandTech used for the 8700K is the Crucial Ballistix Sport then it don't seem to be on the QVL of the motherboard, they also used 4x8 GB there vs 2x8 GB on the Ryzen platform and four memory modules is more demanding for the memory controller so that combined with the motherboard in question may mean they have poor memory performance on the Intel system. They also use a 3U rackmount / small form factor CPU cooler which is just 79 mm high, 88 mm wide which use a 70 mm fan. Then again the Ryzen used the stock cooler so I guess it's completely fair to use a more simplistic cooler for the purpose and they spec what they use but it may not be perfect for the board.

The AMD board mean-while used G.Skill SniperX 2933 MHz and 2x8 GB of them instead.

Anyway the AnandTech result is an outliner. It's not the normal result. The normal result even with these stupidly run gaming benchmarks which may not be tuned for lowest GPU vs CPU load is that the Intel chip win. Sweclockers test use "medium" settings which may lower the demand on the CPU but most likely "definitely" lower the demand on the GPU resulting in more frames and a higher chance the CPU is the limiting factor. And there as you can see above in those games the 8700K is 17-25% faster which is more in line with what you could expect.

The IPC for non-AVX/FPU work-loads seem similar for the two chips but the typical 1 core turbo of the 8700K is 4.7 GHz where's the one thread turbo of the Ryzen 7 2700X seem to end up being 4.35 GHz, the all core turbo of the 8700K is 4.3 GHz whereas the all threads turbo of the Ryzen 7 2700X in some older review was 3.9 GHz but maybe it's closer to 4 GHz on X470? Regardless it's lower. Meaning lower performance / core. Also Intel have 256 bit FPUs vs the 128 bit ones on the Ryzen plus Intel still have a bit lower memory latency for whatever difference that make. The latencies to difference caches and for sharing data between cores varies so that become hard to compare.
In this video: https://www.youtube.com/watch?...
which is great because it show GPU load too you can see that GPU utilization isn't at 100% meaning the CPU is the one holding the performance of these games back but you can also see that the CPU utilization is very low meaning lots of threads / cores are unutilized and performance like are held back by t

Worth noticing is that the 95 watt Ryzen 7 1700X and 1800X increase the power-demands more when going from idle to fully loaded than the 140 watt i7 6900K as well as the whole platform using slightly more power for the whole system:
http://www.sweclockers.com/tes...
So take the TDP values with some salt. It doesn't tell the whole story. Neither have they done for AMD vs Nvidia graphics cards I've heard claimed but there with the advantage for AMD.

Ryzen 7 1700X and 1800X idle whole system at the wall: 45 watt
i7 6900K idle whole system: 60 watt.

1700X blender & x264 benchmark: 155 & 160 watt.
1800X blender & x264 benchmark: 165 & 170 watt
i7 6900K blender & x264 benchmark: 150 & 150 watt.
That's still for the full system.
i7 6900K Intel Ark says 140 watt TDP: https://ark.intel.com/sv/produ...

i7 6950X: 60/145/170 watt idle, blender, x264 (10 core also 140 watt.)

I own Sandy Bridge systems too, and there are some reasons why I am seriously looking at upgrading. A noticeable performance increase, with quite a bit of power saving is always nice. A swedish hardware site has taken up another metric for it, by calculating the amount of joules used in total to render the scene Island in Blender at a set resolution and quality level, minus the rounded off idle energy use by the 980 Ti used in their tests. All results in joule, lower is better:

http://cdn.sweclockers.com/art...

19 September 2014 - Secret press conference - Nvidia will support Adaptive-Sync: http://www.sweclockers.com/nyh...
25 September 2014 - Nvidia won't offer support for Adaptive-Sync but will instead continue with their G-sync and hence never implemented DVI 1.2a support: http://www.nordichardware.se/G...

Both link in Swedish.

Nvidia - Assholes. If true.

They sent it out to some reviewers. Sweclockers has it for example. Google translate it or something if you don't know swedish http://www.sweclockers.com/recension/14699-intel-core-i7-3930k-och-3960x-sandy-bridge-e

I'm disappointed with its power consumption, same as with the Bulldozer. However, unlike with the Bulldozer, you actually get computational performance to match the power consumption.

It's true ATI doesn't have NVIDIA quality drivers; I've never heard of ATI drivers destroying a card. The GTX 590 drivers, however, have a driver issue with their "power limiter" that is supposed to prevent card damaging overvolts. Then there is the whole fan debacle from the 196.75 drivers. I realize people are going say it's what they get for overclocking, but if you add a feature it isn't the user's fault for using it.

[Sources]
http://www.techpowerup.com/reviews/ASUS/GeForce_GTX_590/26.html
http://www.sweclockers.com/artikel/13710-geforce-gtx-590-brinner-i-sweclockers-testlabb-drivrutin-boven-i-dramat [Swedish]
http://ve3d.ign.com/articles/news/53563/Nvidia-196-75-GPU-Drivers-Bugged-Causing-Fans-To-Fail-Cards-To-Overheat

Here's a link to an email conversation with the founder:
http://www.sweclockers.com/forum/showthread.php?s= e0a49aad6aaa76aa2abfe67cb4d4ec44&postid=7539927#po st7539927

The founder also has a column in a small Swedish paper:
http://www.jnytt.se/Read__7981.aspx (Swedish.)
He writes about life in Sao Paulo.

Here's an interview with the largest Swedish business paper, Dagens Industri:
http://tinyurl.com/38ju9l (Swedish.)

A couple of months ago they launched a laptop for 3000SEK (~$450), the same model now being sold for $150:
http://www.idg.se/2.1085/1.100914 (Swedish.)

Could be fraud, could be simply bad business, could be an attempt to gain market share by calculated loss. No matter what, it could definitely have been launched in a more professional way. Bad karma on the founder. I'm waiting (at least) 6 weeks to put in my order.

Some people are really doing their best to research this scam-look-alike. From a swedish website:

----8http://www.cint.se/debatt//ShowPostFlat.aspx? PostID=62603 or http://www.sweclockers.com/forum/showthread.php?s= &threadid=696980 where these letters and answers are mirrored.

"1&2. How could you lower the price so much compared to the price of
3000 SEK that you mentioned in http://www.idg.se/2.1085/1.100914 ?
1000 SEK is quite a bit lower. Why did you change the idea from not selling directly to customers as you staded in http://www.idg.se/2.1085/1.100914

1&2) It is true that we went out in the first article on IDG.se telling
that the price for the end consumer would be around 3000 SEK. That is
because the companies we were discussing with wanted to have a big
profit on each laptop sold to the end consumer. Since we at Medison
believe it is better to offer a laptop so that everyone can afford to
buy a laptop, we decided to offer our Medison Celebrity to a very good
price where we have very, very low margin so that everyone can buy a
laptop rather than giving too high profits to other companies and have
less computers sold to the end consumers."

So the involved companies would just suddenly agree on not making any or have an extremely small margin?

----

"3. Why is this last change and sell-start rushed is such a way? I.E
companys started in just a few days, contracts created in so such
short notice that most of the companies you mention on your site
doesn't know anything about you?

3) There has not been any 'rush', but actually has been carefully
planned over some time. Any kind of registration of company name or
domain does not show wether a strategy is planned or not in a company.
This kind of information can instead be done on purpose to not reveal
to competitors what is going on. But then again, Medison has no
obligations to answer what its strategy is and therefore any kind of
interpretations made by people outside the company."

I would call it either a rush or a failure that the companies listed on your website dodn't know anything about you at first. Which one do you prefer?

----

"4. Why is your company adress sharing the exact same address as
several hundred others?

4) Medison is registered, as many other companies both in UK, USA or
any other European country on an address shared by many companies.
This is quite normal. Let me give you an example. In a high sky scrape
building with 41 floors, all of the people living there have the same
address. We do not see our common shared addres as something strange,
but rather normal."

I wasn't asking about why A company would have it like that. I was asking why YOUR company had it like this. Let me ask the question in another way: Would I correctly assume that your address stated on your site is just a forward or P.O Box like address, and that you don't really have any office in UK at all? If that is the case, do you have any office at all, and what would the visiting address be?

----

"5. Why do a serious company use a hotmail adress to register their
domain? Why is the phone number in the same registration going to
some random person in Malmö and not the registrar?

5) In USA it is VERY common that business people use a hotmail address
or yahoo address together with their business e-mail addresses. It is
common to do so as an insurance to not miss any important e-mails in
case a mail server would crash. Since the companies offering Domain
names have no rules that say a person or a company cannot use a
hotmail address when they register, then we don't see what relevance
your questions has. If the domain company has not updated their client

"If vista scales all the way to 4"

What do you mean? 64-bit Vista scales to 1000 Gigabyte (1TB). 32-bit Vista? Probably just 2 GB.

Here is a warning message "Too little memory" using 2GB with Vista and trying to play "Company of Heroes"

http://img525.imageshack.us/img525/6359/minnetf2.j pg

[I didn't take that screenshot. I saw it linked here:
  http://www.sweclockers.com/forum/showthread.php?th readid=656242%5D

---

To sort my (ehm) porn, I hacked togheter this 8 kb python program using wxPython and pyGame a couple of months ago. Here it is: http://psionicist.online.fr/pile.py.txt

The code is god awful, but it works. Some screenshots here: http://forum.sweclockers.com/showthread.php?s=&thr eadid=504073

Not.
If you take a look in the www.sweclockers.com forums (swedish :P) you'l se lots of people that uses old swedish Volvo car radiators, found at scrapyards. They are the coolest on the forum. They also do lots of other nice stuff. Like Carblaster.

http://www.sweclockers.com/html/artikel/art_040923 _carblaster.php?page=1

Result of Carblaster:
http://www.sweclockers.com/html/artikel/art_040923 _carblaster.php?page=10 //zyk

Not.
If you take a look in the www.sweclockers.com forums (swedish :P) you'l se lots of people that uses old swedish Volvo car radiators, found at scrapyards. They are the coolest on the forum. They also do lots of other nice stuff. Like Carblaster.

http://www.sweclockers.com/html/artikel/art_040923 _carblaster.php?page=1

Result of Carblaster:
http://www.sweclockers.com/html/artikel/art_040923 _carblaster.php?page=10 //zyk

Me and a friend were disassembling an old pentium computer, and were trying to figure out exactly what cpu the machine had. However, there was a lot of really old thermal grease on the cpu, that we just couldn't remove. We tried every solvent we had, but nothing worked. So after a while we got so mad we decided to burn away the grease. We filled a small plate with gasoline, lit it on fire and let it burn for twenty minutes or so. Then we put the cpu back in the motherboard and booted up, and it worked! Weirdly enough, the machine was running an arabic version of Win95, so we couldn't really do much with it.

Pictures of the event is in http://forum.sweclockers.com/showthread.php?s=e7b8 5547f9b2fd539635b02601c5a2ae&threadid=163417&highl ight=eldthis thread. It's in swedish, but there are pretty pictures....

"By the way, I think this is also an expecially American fantasy, a kind of "independence" thing, and that there are very few non-American overclockers."


hmm.. then how come this site has over 30000 registered members and about 1500 individual members logging on every day? and that is in a country with less than half the population of new york...

There is an online report made by Sweclockers.com, Sweden's biggest computer site. The text is in Swedish, but there are tons of photos on case mods etc. Did you ever want to see a wooden case?