Time 15-20min/wu RX 570 mini itx pulse 4GB Polaris 20 XL/Ellesmere clock 1241 Mhz , memory 1500Mhz
AMD Ryzen 5 3600X ratio fix 3.8GHz time 15-35min 12wu/cpu temperatures 60ºC-70º1C Snowman double refrigerator air
motherboard B450 chipset with memory RAM 1600Mhz 2x8gb

You can OC RX clock to 1430Mhz at 1150mv and memory 1800Mhz 0.95v profile timings memory lvl2 time, but you need modificate the fan 100% up to 60ºC.update the drivers adrenalin

actually , no overclock because i need sleep, a lot of sound.

Time 4-8 min/wu RTX4060 OC 8GB Eagle clock 2805Mhz / memory 8250Mhz
AMD Ryzen 9 5900X ratio 42.5 =4.25Ghz time 10-25min 24wu/cpu temperatures 70-75ºC same refrigerator in summer ratio decrease to 42 or less, because , when you up to ratio 42, the temperature increase 10ºC, this is not eficcient (german mind).
Motherboard B550chipset with dual memory RAM 8GBx2 at 3600mhz

You can overclok rtx4060OC +1500Mhz memory and +195Mhz the clock, if you use undervolt you can make stable +1500Mhz memory = 9750mhz and +195Mhz clock = 3000Mhz at 1050mv or 3045mhz at 1075mv or 3075mhz at 1100mv, but , more tha 3045mhz the clock ...you need increase mhz of memory more tan 9750mhz. The temperatures in RTX4060 are very very low. 43ºC , this wu, no use much GB memory.

Actually, no overclock, because, i see, in time, that no is eficcient, no for faisl of memory in the wu, because no error wit the progma analyser HWiNFO64, in sensors, last of list --> erros gpu = 0.

its no efficient because the times no apreciate less time per wu

Now i Buy AMD Ryzen 9 9950X
and i buy a RX 7900XT, and RX9070XT,
motherboard chipset X870P..
memory 2x8Gb 6000Mhz DDR5, BUT!!!,,,the cpu only can 5600mhz in the nexts days
i will publicate the results

thx probably the rx9070xt doesnt work, beacuse is very recently architecture navi 48, but , its practically the same at rx7900xt, the differente is, the nwe rx9070 have ray tracing but 4GB less and less power of floating points

I dont know how many wu per day can make the nvidia rtx 4070 vs amd rx7900xt,

Ray tracing improvements aren’t relevant for crunching. The new generation is expected to be slightly more efficient, but I think the 9070 XT and 7900 XT should deliver similar performance.

Both cards should already be supported. Let us know once you have some results to share.

Ray tracing improvements aren’t relevant for crunching. The new generation is expected to be slightly more efficient, but I think the 9070 XT and 7900 XT should deliver similar performance.

Both cards should already be supported. Let us know once you have some results to share.

thank you for reply this message anorek , efectively,

RX7900XT is ..better than rx9070xt, 460 GPixel/s vs 380.2 GPixel/s

The rx9070xt only works better in games because have ray tracing, only is make for competitive actually in nvidia, but, nvidia , no have competitors, because nvidia have DDR7 and much more GB memory.

Probably is interesting, no buy the new rx9070xt, and buy rtx5000 series, by , in a other hand... without sli cross fire, ¿can run two graphic cards two wu isolate in the same motherboard?, because the motherboards have 1 pcie 5.0 and two 4.0, totally 3 pcie

please delete this new reply because im RETARD, i generate accidentally when was writing the quote

https://asteroidsathome.net/boinc/show_host_detail.php?hostid=773485

host with rx 9070хт. If you are interested in evaluating the operating time.

https://asteroidsathome.net/boinc/show_host_detail.php?hostid=773485

host with rx 9070хт. If you are interested in evaluating the operating time.

¿how can evaluate?, is off ¿no?, its a few points, ¿its run 24h/day?

the rx9070xt make task in 200 to 400seg, same as RTX4060, mmm, this is not possible! OMG

I have another quuestion, ¿what is grc pool?, is asteroids?

I have another question, What does it affect, for the processing of the tasks of the graphics card that?
is 5x PCI-E x16 slot

PCI_E1 Gen PCIe 4.0 supports up to x16 (From CPU)
PCI_E2 Gen PCIe 3.0 supports up to x1 (From Chipset)
PCI_E3 Gen PCIe 3.0 supports up to x1 (From Chipset)
PCI_E4 Gen PCIe 3.0 supports up to x4 (From Chipset)
PCI_E5 Gen PCIe 3.0 supports up to x1 (From Chipset)

if you have PCIe 3.0 is slow, thna 4.0, but, internally, the graphic works on his memory, that is important, obtain PCIe5.0?

¿Its important de velocity of PCIex16 on 3.0 vs 5.0 in asteroids at home with rx7900xt?

I have Psi-e 4 set in BIOS because the motherboard supports the fourth version. I myself am unpleasantly surprised by such performance.

I have Psi-e 4 set in BIOS because the motherboard supports the fourth version. I myself am unpleasantly surprised by such performance.

I think, i cant run at 100% rx7900xt because the second slot works at x8 if conect first slot, and works two cards in Asus prime x870p. I cant pay motherboard workstation with epyc cpu.

I have Psi-e 4 set in BIOS because the motherboard supports the fourth version. I myself am unpleasantly surprised by such performance.

I think, i cant run at 100% rx7900xt because the second slot works at x8 if conect first slot, and works two cards in Asus prime x870p. I cant pay motherboard workstation with epyc cpu.

¿In this tasks of BOINC are really relevant obtain 5.0 to 4.0 or, x16 or x8? , much motherboards only works at 8x when are two graphic cards connected at motherboard, ¿this is relevan in this type of tasks?

Example actually have motherboard asus prime B550 whit RTX4060 OC 8GB Eagle, and...the motherboard have 1 PCIe 4.0 at x16, and actually en 2 weeks have 20.000 promedio, in fact, i make less more than 30.000

My mother board Placa Base ASUS PRIME B550M-A WIFI II . have

PCI Express x1 (Gen 3.x) ranuras: 2
PCI Express x16 Gen (3.x) ranuras: 1
Ranuras PCI Express x16 (Gen 4.x): 1



For example, using the HWINFO64 , i see Bus PCIe 4.0 X8 16GT/s and x8.

The velocity of RAM are 4.6Gb/s

https://ibb.co/Ps5jjscc

theoretical gigatrasnfers per second rx7900xt

Pixel Rate
459.6 GPixel/s

Texture Rate
804.4 GTexel/s

PCIe bandwidth largely does not matter. you will get the same speed with PCIe 3.0 x4 as with PCIe 5.0 x16, it doesnt matter for the applications here.

GPU specs like pixel fill rate and texture rate also do not matter and have nothing to do with how the app here operates.

the app here cares more about memory bandwidth and clock speed.

---

ok, thx, but its the same x8 than x1? because the motherboard its turn to x1 if you connect two graphic cards

PCIe bandwidth is mostly important for SSD performance or when games exceed available GPU memory. In typical scenarios, it's primarily used during data transfers to the GPU, like when loading a game level. Compute tasks, on the other hand, spend minimal time on data loading. Most of their time is spent processing data that's already resident in GPU memory. Faster is always better in theory, but if data transfers only account for 1% of the total workload, speeding them up even 10 times won’t make a noticeable difference.

what doesn't matter:
* pcie bandwidth
* tensor / AI / wmma - no boinc app utilize them
* FP16, INT4... - despite great perf numbers, no boinc app utilize them
* TMUs (TMU * clock speed = Texture Rate) - only relevant for games
* ROPs (ROP * clock speed = Pixel Rate) - only relevant for games
* VRAM capacity - A larger size won’t improve performance, and most GPUs have more than enough capacity to handle BOINC projects.

what matters:
* GPU vendor / architecture in general
* cores - but meaningful comparisons can only be made within the same architecture and vendor
* FP32
* FP64
* INT32
* cache
* memory bandwidth
* clock speed

Those are attributes you should be looking for, but just like with games, performance can vary. Some projects see greater gains from high clock speeds, while others rely more on memory bandwidth. It depends on the type of computations and how well the application is optimized.

Note that FP32 represents peak performance under ideal (unrealistic) conditions. Real-world applications are more complex than simply multiplying two numbers. As a result, newer architectures can outperform older GPUs in certain applications, even if their FP32 performance is lower because they can utilize available resources more efficiently.
The numbers can give you a rough performance estimate, but it's always best to test with the specific application you plan to run.

Note that FP32 represents peak performance under ideal (unrealistic) conditions. Real-world applications are more complex than simply multiplying two numbers. As a result, newer architectures can outperform older GPUs in certain applications, even if their FP32 performance is lower because they can utilize available resources more efficiently.
The numbers can give you a rough performance estimate, but it's always best to test with the specific application you plan to run.

I buy 2 rx7900xt 2x789€ motherboard B840 and 1 nvidia rtx5070ti OC 949€ motherboards B850

In 1 month will recibe the cpus Ryzen 9 7950x :)

Note that FP32 represents peak performance under ideal (unrealistic) conditions. Real-world applications are more complex than simply multiplying two numbers. As a result, newer architectures can outperform older GPUs in certain applications, even if their FP32 performance is lower because they can utilize available resources more efficiently.
The numbers can give you a rough performance estimate, but it's always best to test with the specific application you plan to run.

I buy 2 rx7900xt 2x789€ motherboard B840 and 1 nvidia rtx5070ti OC 949€ motherboards B850

In 1 month will recibe the cpus Ryzen 9 7950x :)

¿Can you use the IA tensor cores of nvidia to program ASIC of asteroids at home?

> Can you use the IA tensor cores of nvidia to program ASIC of asteroids at home?

No, tensor cores do support low-precision data types like FP16, they're specifically built for AI workloads. The Asteroids app depends on FP64, and lower precisions aren't good enough. That’s also why none of the existing BOINC projects make use of tensor cores (or NPUs); they're not suitable for scientific computing. Like ASICs, tensor cores are highly specialized, great for AI, but pretty much useless for most scientific applications.

> Can you use the IA tensor cores of nvidia to program ASIC of asteroids at home?

No, tensor cores do support low-precision data types like FP16, they're specifically built for AI workloads. The Asteroids app depends on FP64, and lower precisions aren't good enough. That’s also why none of the existing BOINC projects make use of tensor cores (or NPUs); they're not suitable for scientific computing. Like ASICs, tensor cores are highly specialized, great for AI, but pretty much useless for most scientific applications.

Lo probaré

¿Do you recommend crossfire/SLI or not,?

¿Do you recommend crossfire/SLI or not,?

no, it's pointless for any BOINC project. it will not do anything.

---

PCIe bandwidth is mostly important for SSD performance or when games exceed available GPU memory. In typical scenarios, it's primarily used during data transfers to the GPU, like when loading a game level. Compute tasks, on the other hand, spend minimal time on data loading. Most of their time is spent processing data that's already resident in GPU memory. Faster is always better in theory, but if data transfers only account for 1% of the total workload, speeding them up even 10 times won’t make a noticeable difference.

what doesn't matter:
* pcie bandwidth
* tensor / AI / wmma - no boinc app utilize them
* FP16, INT4... - despite great perf numbers, no boinc app utilize them
* TMUs (TMU * clock speed = Texture Rate) - only relevant for games
* ROPs (ROP * clock speed = Pixel Rate) - only relevant for games
* VRAM capacity - A larger size won’t improve performance, and most GPUs have more than enough capacity to handle BOINC projects.

what matters:
* GPU vendor / architecture in general
* cores - but meaningful comparisons can only be made within the same architecture and vendor
* FP32
* FP64
* INT32
* cache
* memory bandwidth
* clock speed

Those are attributes you should be looking for, but just like with games, performance can vary. Some projects see greater gains from high clock speeds, while others rely more on memory bandwidth. It depends on the type of computations and how well the application is optimized.

The RTX5070Ti have 336 tensore cores on FP64 , chip GB203 Blackwell architecture

https://asteroidsathome.net/boinc/forum_thread.php?id=1125