[BitsAndChips]390X ready for launch - AMD ironing out drivers - Computex launch

[Silverforce11]

I just bought a 295x2 because it was so cheap and wanted to check out the AIO. Definitely not going with the WCE. Runs cool and quiet but raises the temps on all my other components due to exhaust restrictions. Losing that 140mm rear fan has a huge deficit under idle conditions.

If you use it as an exhaust, why does it raise ambient temps?

An exhausting radiator with good CFM is doing two job, removing heat from the GPUs as well as extracting ambient air from the case. Thus, its akin to having a dedicated exhaust fan. Unless you are using it as intake?

 

[guskline]

If you use it as an exhaust, why does it raise ambient temps?

An exhausting radiator with good CFM is doing two job, removing heat from the GPUs as well as extracting ambient air from the case. Thus, its akin to having a dedicated exhaust fan. Unless you are using it as intake?

With the 295x2, since it is an AIO, it still has a fan to cool the vrms, the heatsinks only works on both gpus. Perhaps the heat of the vrms raises everything even at idle. That vrm fan is throwing that heat out into the case.

In my setup with 2 r9-290s with EK custom waterblocks, BOTH the gpu and vrm are transferring heat to the waterblock, then into the water loop and then into the rad and out. Little if any vrm/gpu heat is dissipating inside the case.

If the Rad for the AIO replaces an unimpeded 140mm exhaust fan I can see how that would raise overall internal temps especially with the vrm heat.

From what I read, this AIO is not bad but is designed to keep only the 2 290x gpus from throttling.

 

[dacostafilipe]

If you use it as an exhaust, why does it raise ambient temps?

An exhausting radiator with good CFM is doing two job, removing heat from the GPUs as well as extracting ambient air from the case. Thus, its akin to having a dedicated exhaust fan. Unless you are using it as intake?

When I added my CPU AIO, I had to adjust the air flow to.

Maybe Rvenger should look into that, maybe adding the 140mm fan on top/bottom of his case.

 

[Cloudfire777]

EDIT: Just to clarify, the original source of this slide is NOT Chiphell. Its the same guy that gave us the information about the card I wrote about below
Here is the original picture without Chiphell logo
Original source here: http://tieba.baidu.com/p/3672666940?pn=1

I really like this picture

I found some information about the card
Translated by google, and freshened up by me.

According to industry veterant, famous editor Hao, AMD HQ have confirmed that the 390X is really ready for launch. It is reported that the 390X will be using GCN 1.3 architecture, have full support for DX12, Freesync and improved color frame buffer compression technology. This flagship edition is devided in to air cooled and a water cooled version. The first step will be to release the water cooled version, top turbo clocks all the way up to 1200MHz, 8GB HBM memory and the opponent will be GTX Titan X. Air cooled version will later battle out with GTX 980 Ti, unknown specifications at this point.
We also learned that this card (Water cooled edition) is expected to sell for $899, TDP is 250W, outstanding value for money.

Source: http://tieba.baidu.com/p/3672666940?pn=1

 

[csbin]

Fool's Day

 

[Noctifer616]

Any other day this might be exciting, but not today.

 

[Gloomy]

Not today.

Though if it really is 1200mhz then it would explain the 300w+ rumors... gcn hates clocking over 900mhz.

 

[Ranulf]

Legit or not, I really need to learn to just shut off the internet on April 1st. Even Amazon has a silly joke front page.

 

[dacostafilipe]

Hunting Titans? Radeon Olympus?

 

[el etro]

If fake, the slide is well done. Very well done.

 

[Cloudfire777]

If fake, the slide is well done. Very well done.

I`d say this is a fake one

 

[Madpacket]

Latency is very important to keeping shaders fed. I tested lots of vram settings during the bit mining craze, even lowering vram clocks but decreasing latency resulted in massive throughput. It was totally against what we back then were expecting, everyone pushed vram OCs to the max but they found that typically the performance did not improve unless the clock speed fell into some ranges that reduced latency.

Definitely an entire new memory subsystem to move things around within the GPU requires new drivers taking advantage of that.

Can confirm this myself.

 

[AdamK47]

April 1st. There's no way they could beat the Titan X.

 

[stahlhart]

Merged to already existing R9 390X discussion.
-- stahlhart

 

[Rvenger]

If you use it as an exhaust, why does it raise ambient temps?

An exhausting radiator with good CFM is doing two job, removing heat from the GPUs as well as extracting ambient air from the case. Thus, its akin to having a dedicated exhaust fan. Unless you are using it as intake?

Because I used my last exhaust fan outlet. When the card is idle the fan hardly pushes air out the back of the case thus impeding the airflow and causing my idle temps on my CPU to jump over 40c, then all my fans start increasing to keep up with the heat. I believe it's the VRMs on the 295x2 that contribute to the extra heat. Not all the heat is dispelled in the radiator.

 

[5150Joker]

Lol ruby is looking old and fat. Maybe she's pregnant, I'd put her at 28 weeks.

 

[VulgarDisplay]

April 1st. There's no way they could beat the Titan X.

I'm just curious why not?

 

[Rvenger]

Are you suggesting to rebrand Hawaii (R9 290X and 290) and at the same time regress performance as well with lower clocks ??

Not going to happen.

Yes, it becomes a much more efficient chip that way. They need to cut down Fiji another tier in order to make it possible. 300w Hawaii is not something we want when Fiji will probably be the same TDP.

 

[Cloudfire777]

Merged to already existing R9 390X discussion.
-- stahlhart

You sure know how to hide stuff from exposure in this forum.

Lol ruby is looking old and fat. Maybe she's pregnant, I'd put her at 28 weeks.

hahaha.
Its the inside (hardware) that counts Joker

Moderator callouts are not allowed, even with a smilie face. Responses to a moderator's actions should only be discussed in the moderation forum or through direct PMs.
-Subyman

 

[NTMBK]

April 1st. There's no way they could beat the Titan X.

You think the move to a massively better memory technology won't make any difference? Just go look at NVidia's slides about Pascal...

 

[stahlhart]

You sure know how to hide stuff from exposure in this forum.

We don't need duplicate rumor discussions here about the same card, nor do we need April Fool's jokes about said rumors.
-- stahlhart

 

[Cloudfire777]

You think the move to a massively better memory technology won't make any difference? Just go look at NVidia's slides about Pascal...

If you are referring to this picture, stacked memory is only a part of the equation. Much faster communication between CPU and GPU through NVLink plus better scaling from GPU to GPU and 4x more FP16 cores are just as important.

HBM will undoubtly give some boost but it also depends on the engine behind it all, ie the cores. It will be interesting to see what AMD have done

 

[Cloudfire777]

Pretty interesting that a reverse google search only found AMD slides
The style looks exactly like AMD use. Maybe its not fake?

https://www.google.no/search?tbs=sb...a=X&ei=dfQbVefMC4G4sAGf5ILYAg&ved=0CBkQ9Q8oAA

 

[ShintaiDK]

You think the move to a massively better memory technology won't make any difference? Just go look at NVidia's slides about Pascal...

Problem is HBM1 isnt that. HBM2 however is.

 

[ShintaiDK]

Pretty interesting that a reverse google search only found AMD slides
The style looks exactly like AMD use. Maybe its not fake?

https://www.google.no/search?tbs=sb...a=X&ei=dfQbVefMC4G4sAGf5ILYAg&ved=0CBkQ9Q8oAA

People would never photoshop from something existing...oh wait.