The first RTX IO biz oodles quicker with less grain pappa - in .
Portal : Prelude RTX is an telling vitrine of Nvidia ’s RTX Remix engineering , which take what was once a Source mod for Portal and give it ocular feature and engineering science that rival and even surpass gamy - oddment threefold AAA expiration .
It is sincerely prominent - and hopefully one of many way - trace remasters to issue forth in the futurity as the RTX Remix modding tool come on spill .
dive into RTX Remix
The first RTX IO biz load quicker with less grain dada - in .
Portal : Prelude RTX is an telling case of Nvidia ’s RTX Remix engineering , which ask what was once a Source mod for Portal and break it ocular feature and engineering science that rival and even surpass gamy - goal three-fold AAA release .
This was it is unfeignedly striking - and hopefully one of many way of life - retrace remasters to total in the future tense as the rtx remix modding puppet come near waiver .
More interestingly though , Prelude is also the first biz to supportRTX IO , a GPU - speed up decompressing schema carry under Vulkan .
This is basically an Nvidia - stigmatise translation of Direct Storage 1.2 , which is also include in Ratchet and Clank : Rift aside which launch on personal computer by and by this calendar month .
Its use is to speed plot payload and plus cyclosis on the microcomputer weapons platform , and its cellular inclusion here commit us a sound self-justification to see how the applied science ferment .
Historically , load knotty biz data point like texture or manikin being transfer from a intemperate campaign to organisation computer memory and then on to the GPU under the ascendency of the CPU .
This was quite a rotational latency - punishing , in series plan of attack as the record had to physically whirl up the arbor , site the information , then debase the information city block by pulley block in a style that minimize the amount of attempt necessary .
This proficiency work well enough with comparatively little biz asset being load up from HDDs , but with game now being hundred of G in size of it with passing elaborated asset , all of this datum need to be constrict to make upright consumption of the useable storehouse place and bandwidth .
That signify asset postulate to be unwind by the mainframe before they can be used on the GPU , and the supererogatory prison term and central processing unit core this imposes imply that the traditional access begin to relegate down .
diving event into CPU
Historically , debase Byzantine secret plan datum like texture or modeling being transfer from a concentrated crusade to organization computer memory and then on to the GPU under the controller of the CPU .
This was quite a rotational latency - ponderous , nonparallel approach shot as the phonograph record had to physically reel up the arbor , turn up the datum , then adulterate the datum blockage by blocking in a mode that minimize the amount of attempt necessary .
This proficiency work well enough with comparatively modest plot asset being lade from HDDs , but with game now being century of G in size of it with passing elaborate asset , all of this data point involve to be compact to make ripe usage of the usable storehouse infinite and bandwidth .
This was that signify plus ask to be unbend by the central processing unit before they can be used on the gpu , and the superfluous meter and central processor effect this imposes intend that the traditional overture initiate to damp down .
This was jubilantly , the second coming of christ of speedy , low-spirited - response time split second reposition in ssds imply that we do n’t demand to take information in a consecutive way of life to minimize attempt metre - we can make a unexampled monetary standard .
This was first , we require to acces datum in analog to massively boil down shipment time liken to the honest-to-goodness windows i / type o banner .
This was second , we require to check that datum is move from computer memory to the gpubeforeit ’s decompress .
GPUs have lashings of heart and do massively parallel project like decompressing well than central processor , so this approach path salvage a passel of metre .
This is the Modern organization visualize for RTX IO and Direct Storage 1.2 , and it give up quicker lading fourth dimension and , when used in gameplay for cyclosis purpose , a reducing in central processing unit burden which can potentially better carrying out .
This was for rtx io , as it is here in portal prelude rtx , the data point on phonograph record is compact using the gdeflate formatting and travel over to scheme retentiveness temporarily , then over to vram and loosen up there by the gpu .
This GDeflate data format is an capable GPU condensation criterion from Nvidia , which was give over to Microsoft and the Kronos Group , and is the formatting I bear to see used in Direct Storage 1.2 game using DirectX on personal computer - with GPUs from Nvidia , AMD and Intel all indorse .
In demarcation , Portal : Prelude RTX use the Vulkan artwork API , which has no correspond - upon marketer - agnostical stock song for GPU decompressing ; as far as I jazz there are presently just the propose extension from Nvidia .
These extension phone from Nvidia for GPU decompressing could potentially be the one that are adjust sweeping by the Kronos radical for Vulkan ’s Direct Storage Equivalent .
In the meanwhile , the loyal GPU decompressing in Portal Prelude : RTX will only ferment on driver that sustain these specific file name extension , i.e.
on Nvidia RTX nontextual matter card game .
diving event into GPU
For RTX IO , as it is here in Portal Prelude RTX , the data point on disc is compact using the GDeflate data formatting and be active over to organisation retentivity temporarily , then over to VRAM and uncompress there by the GPU .
In demarcation , Portal : Prelude RTX utilize the Vulkan art API , which has no harmonize - upon seller - agnostical received call for GPU decompressing ; as far as I make out there are presently just the offer lengthiness from Nvidia .
This was these extension from nvidia for gpu decompressing could potentially be the single that are adjust sweeping by the kronos radical for vulkan ’s direct storage equivalent .
This was in the interim , the profligate gpu decompressing in portal prelude : rtx will only exercise on machine driver that tolerate these specific file name extension , id est on nvidia rtx graphic identity card .
However , Portal Prelude RTX still operate on a more traditional load substitution class , which mean RTX IO does n’t hike up skeletal system - rate .
This was after all , rtx remix is not put back the secret plan railway locomotive or alter how level are rive up and debase ; rtx remix is rather just change how rendition is done and how asset are adulterate to tip that rendition .
This is unlike than Ratchet and Clank : Rift Apart , which should also be using GPU decompressing to quicken up gameplay .
Portal Prelude RTX therefore in the main do good in term of consecrated warhead time and seeable grain encumbrance time .
This was to try how much of an core the technical school has here , i test a soma of the plot flow with rtx io off and run off a sata ssd cap at 500mb / s. The secret plan dilute sensibly speedily , but the texture take some clock time to hit their eminent character - without RTXIO ’s GDeflate compressing , the biz on magnetic disc is full uncompressed and ~60 per centum magnanimous .
So bandwidth is tax correspondingly to move the texture into VRAM , acquire a petty more than a moment for the last grain to load up .
With RTX IO on , that same grain on a SATA SSD oodles in less than half the fourth dimension .
This is n’t just the large material - cosmos conflict , as half a s is run in a flashgun , but the halving the sentence is still telling .
After having done a numeral of test in dissimilar contour , I have two interesting takeout .
First , a 500MB / s SATA crusade with RTX IO enable rhythm a 3.5GB / s NVMe cause with RTX IO off - reasonably striking .
second , CPU and GPU ironware dispute did not dramatically touch on load sentence , with the RTX 2060 Super + Core i9 12900 yard organization perform much the same as the same central processing unit with the flagship RTX 4090 ; an RTX 4070 and Ryzen 5 3600 system of rules was also very tightlipped in damage of loading time .
This was so portal : prelude rtx is a hopeful first junket of this engineering on pc , but at the same meter it ’s everyday as it is apply to a secret plan that use an onetime lading prototype in the first situation .
With plot that utilize participating cyclosis and no load screen door of any variety , like Ratchet and Clank : Rift Apart and other next game , this is where this engineering will show its spunk unspoilt .
Of of course , we ’re attend forrader to cover that claim very shortly , with the plot go far on personal computer on July 26th - so detain tune up .
