EPTEdit: I'm liking the AMD Smart Shift technology. That could be something pretty huge for laptops if the numbers check out.
Computer Build, Upgrade & Buying Resource Thread - Page 709
| Forum Index > Tech Support |
When using this resource, please read the opening post. The Tech Support forum regulars have helped create countless of desktop systems without any compensation. The least you can do is provide all of the information required for them to help you properly. | ||
|
ZerOCoolSC2
9064 Posts
Edit: I'm liking the AMD Smart Shift technology. That could be something pretty huge for laptops if the numbers check out. | ||
|
Cyro
United Kingdom20343 Posts
The zen2 APU's in general will be enormous for laptops. Hilarious that a 15w zen2 system would have 4% faster ST and 90% faster MT vs a 15-25w icelake system - faster cores AND twice as many of them, plus a frequency/voltage curve that's super friendly to widening the workload. They showed off for context how bad zen1/zen+ was in laptop parts by comparison with a claimed 2.24x increase in perf/watt, mostly from the 7nm process. | ||
|
ZerOCoolSC2
9064 Posts
On January 07 2020 07:51 Cyro wrote: It's pretty basic, but yeah 10% boost sometimes. In a mixed cpu/GPU workload it might be relatively power-free but in the cinebench example they gave, that 12% performance uplift WILL have cost power. It doesn't change peak power though, only average power - and that's beneficial for designs limited by peak power. The zen2 APU's in general will be enormous for laptops. Hilarious that a 15w zen2 system would have 4% faster ST and 90% faster MT vs a 15-25w icelake system - faster cores AND twice as many of them, plus a frequency/voltage curve that's super friendly to widening the workload. They showed off for context how bad zen1/zen+ was in laptop parts by comparison with a claimed 2.24x increase in perf/watt, mostly from the 7nm process. I know in architecture, a lot of students use laptops to do renders or get presentation boards done. I think that would be something that could be definitely useful to them and other design oriented careers. I hated having to shut down other programs to run one because there just wasn't enough power to go around. (Granted, I was running at most 8GB at the time). | ||
|
Cyro
United Kingdom20343 Posts
| ||
|
ZerOCoolSC2
9064 Posts
Thanks for the clarification. | ||
|
Cyro
United Kingdom20343 Posts
7.14x performance per dollar and with the reduced complexity of a single-socket board murder streamed live on youtube | ||
|
ZerOCoolSC2
9064 Posts
On January 07 2020 08:13 Cyro wrote: Also lol at threadripper. $4000 64c128t CPU outperforming the $20,000 flagship dual-28c56t Intel system by 43%. 7.14x performance per dollar and with the reduced complexity of a single-socket board murder streamed live on youtube RIP in peace hahaha. That was a highlight indeed | ||
|
Cyro
United Kingdom20343 Posts
Very stable frametimes. It can only be set to whole framerate intervals, unlike RTSS which lets you specify frametime in fractions of an FPS or even microseconds. RTSS displays frametimes in tenths of a millisecond, so i can RTSS lock to exactly 4.3ms and it'll require a 0.1ms+ spike to show up as anything other than 4.3ms on the monitor. For example, a 4.3ms frametime is 232.558fps. The Nvidia driver can't set that, it can only lock to 232.0 or to 233.0fps which makes my monitoring tool (detecting swings of up to 0.1ms) less accurate. It still seems to be very similarly stable though, waaaaaaay more stable than an uncapped FPS would be. --- Latency is low - how low exactly, i can't say; that's a great indicator for it performing well, at least. The 1 frame lag of RTSS at 240hz isn't very perceptible, particularly because it's constant. It's hard to tell the difference between an ultra-low, stable input lag and an extremely-ultra-low, stable input lag. | ||
|
Dingodile
4139 Posts
| ||
|
Cyro
United Kingdom20343 Posts
Still.. limiting to 30 has got to be horrific in all sorts of ways, especially through an out-of-engine limit. | ||
|
FiWiFaKi
Canada9859 Posts
I have this case, and I'm wanting to mount a 280mm radiator, in the reviews, nobody has mounted it this way, the case doesn't mention it, but I don't see why not. Top top of the case has a very fine mesh (8 hexagons per linear inch), then there's a gap of at least 25mm (looks like 30mm to me, I didn't measure it), but enough to fit a standard 25mm fan, and afterwards there's a course mesh (3.5 hexagons per linear inch), this course mesh has mounting areas for 2x120mm fans or 2x140mm fans. Below the mesh there's a gap of roughly 32+mm to the VRM's, and 40mm to the RAM (a 120mm wide radiator would go above the VRM's because the mounting location is offset towards the glass panel), a 140mm wide radiator would not clear the VRM's, so let's say I have 32mm of space below the mesh to work with. Would it be an effective solution to mount 2x140mm fans between the two meshes, and then mount a 280mm radiator below the course mesh (in essence having a mesh between the radiator and the fans), I should also point out that at the mounting points, they indent 1-2mm into the middle of the case, so the fans would be tight against the mesh, but the radiator would be 1-2mm from the surface of the mesh. Unsure how large a concern not having the fans against the radiator would be, less blockages on the sides might make air seep between that gap, rather than force its way through the blockages in the radiator. There's probably 2" upfront from the mounting area, and 1.5" back, and the holes are around 1" slits so I could move the radiator to make it fit I think. Would this work? I like the idea of having things a bit cramped, makes it feel nice and cozy, plus I quite like how my build looks here, and for once I wouldn't mind going watercooling to have it more aesthetically pleasing... Or should I just get a Noctua NHD15S or Dark Pro 4 (I'd have to check compatibility, of course I bought Spectrix D60G RAM which is 46mm tall lol). Thanks in advance! | ||
|
Cyro
United Kingdom20343 Posts
Unsure how large a concern not having the fans against the radiator would be That's a pretty big problem. Radiators are tightly finned and reliant on static pressure to move air through them, any gap and you're now bouncing a bunch of air off the radiator rather than forcing it through. --- Here we go, video i was waiting for :D Basically it's nearly identical performance to RTSS, less configurable but less likely to have compatibility issues. I'l use RTSS since i don't have any compatibility issues at the moment. | ||
|
FiWiFaKi
Canada9859 Posts
A high static pressure fan creates a pressure of roughly 0.02kPa... So that's 0.02% of atmospheric pressure... 20 Pascals * 0.04m^2 (area of a 280mm fan) = 0.8N... So 80g of force imparted on the fan if all of it was lost (but say you lose 50% of the static pressure going through the radiator), the fans are only applying a 40g force on the radiator. I feel kind of stupid as someone with a mechanical engineering background not knowing how to analyze a fan, we just always dealt with closed loop systems, and usually the static pressure pressure is much greater than the dynamic pressure (density of the air x velocity^2). But this case, to me seems sort of like a waterwheel being impacted by water, sure the water can go around the blades of the wheel, but its high kinetic energy pushes water through turning the wheel. My head says that by having a hole between the fan and radiator that is 2mm at most would have only a tiny detriment of airflow under 5%. The holes in a radiator seem pretty big, when I blow at a radiator, plenty of air goes through, rather than going around. As a rough back of the envelope calculation: Air flow through a 140mm fan attached to a radiator, 60m^3/h -> 0.166m^3/s, the surface area of the fan where air goes through is roughly half the area of the 140mm square, and lets assume it has the same velocity everywhere. So 0.166m^3/s / (0.14m*0.14m) -> v_avg = 8.5m/s q = 0.5*density*v^2 q = 0.5 * 1.2kg*m^-3 * (8.5m/s)^2 q = 43 Pascals So the dynamic pressure is over twice the static pressure (and it's the static pressure that would be responsible for the air seeping through the cracks. Either way, I could purchase some backer rod, or make a gasket if I don't want to be concerned with that. | ||
|
Broetchenholer
Germany1967 Posts
She's got: https://www.arlt.com/pc/gaming-pcs/allround/arlt-gamer-amd-ryzen-5-8gb-ssd-radeon-vega-11-arlt-1.html I am thinking about putting 8gb more memory in or buying a dedicated graphics card, but I am not sure which is better. Not sure if a gx card for less then 100 euros would make that much of a difference. | ||
|
ZerOCoolSC2
9064 Posts
| ||
|
Broetchenholer
Germany1967 Posts
| ||
|
Lmui
Canada6223 Posts
You'd be best off getting something siimple like the 1650 super - https://www.mindfactory.de/search_result.php?select_search=0&search_query=1650 super It requires a 6 pin power connector though, and I'm not sure it has one (check inside for a connector off the power supply that looks like the black connector in the picture: ![[image loading]](https://images-na.ssl-images-amazon.com/images/I/61r0HPMOxYL._AC_SL1024_.jpg) If you have one, you'll be good. The 3400g isn't super powerful but it should easily be enough for 60+fps in most cases. edit:: https://www.techspot.com/article/1532-pubg-cpu-benchmarks/ The 3400g is in the neighbourhood of the 1500X-1600x, so should comfortably go above 60 with a GPU. | ||
|
Broetchenholer
Germany1967 Posts
| ||
|
Lmui
Canada6223 Posts
On January 15 2020 02:45 Broetchenholer wrote: That's too much for a present, she locked me into less than 100 euros due to my super expensive Christmas gift. Any chance you two could split the bill for this (Let her know that it's probably the best value upgrade she could get?). The GTX 1050 TI is the only new part that you can get at a lower price (Still 140 euros or so for a pretty old card at this point). If you're willing to try 2nd hand, the RX 480/RX 570 can be found on ebay for fairly low prices (under 100 euros). The only caveat is that these are going to be used cards, that were likely run 24/7 for months/years while mining, but if you get a good one, they're the best value cards you can get on the market anywhere. Same caveat about the power connector applies. | ||
|
Broetchenholer
Germany1967 Posts
| ||
| ||
