How hot the earth will be if we have 20% CO2 in the atmosphere instead of ~0.04? and how cold would it be if doesn't have any CO2?

The concentration of CO₂ in the Earth's atmosphere has remained less than 1% in the past 470 million years when land plants first appeared--long before mammals, and long before humans graced the earth.

If we are talking about having a biosphere comparable to the one we know, we need to be talking about CO₂ and global average temperatures that are comparable to levels during the relatively stable Holocene epoch of the past 11,000+ years. Those are the conditions that support life, with the current overshoot beyond 350 ppm being brief (in terms of geological time).

As for 0% CO₂ concentration in the Earth's atmosphere, that would remove warming that essential for life on the planet. The goal is to have CO₂ levels that are stable and also at a level which provides the warming for which life in the biosphere is accustomed to--which is less than 350 and probably even less than 325 ppm. An accepted scientific paper is Target Atmospheric CO₂: Where should humanity aim? [https://arxiv.org/pdf/0804.1126.pdf].

This discussion should also switch from the concentration of CO₂ in the atmosphere to global emissions from a few human activities (roughly 90% being combustion of fossil fuels). The question to ask is how much do those human emissions need to change for the current annual increases in CO₂ concentrations to stop rising? They need to return to emission levels that were essentially zero as they were prior to the Industrial Revolution. That is when natural global emissions and absorption (aka sinks) cancelled one another out to keep levels stable. However, cutting global human CO₂ emissions by half (or let's say 60% to be safe) will stop the rise in CO₂ levels for a few years. If cuts do not go further within the decade toward 100%, CO₂ levels in the atmosphere will start to rise again. Reference: IPCC AR4 FAQ 10.3 [https://archive.ipcc.ch/publications_and_data/ar4/wg1/en/faq-10-3.html].

There are other greenhouse gasses besides carbon dioxide (in particular water vapor) but in order to give a definite answer, let's pretend the atmosphere is completely free of them. In this case, it's easy to calculate the average temperature of the earth by holding the incoming shortwave heat flux at its present value and assuming that it is balanced by outgoing blackbody radiation from the surface.

Similarly, rather than deal with your 20% carbon dioxide atmosphere explicitly, let's just say the atmosphere provides a "perfect" single-layer greenhouse blanket: an isothermal one that allows shortwave radiation through in the same way the existent one does, but absorbs all the outgoing longwave radiation (I don't think anywhere close to 20% concentration of carbon dioxide would be required to approach this limit). This greenhouse blanket reaches a temperature such that it radiates the same amount of heat to space as is coming in from the sun. Since it radiates equally upward and downward (and we assume all of the downward longwave radiation is absorbed), this means the total energy absorbed (and therefore, in equilibrium, radiated) at the surface is doubled relative to what it would be with no greenhouse blanket. Note that a larger warming effect is possible if the “blanket” is allowed to stratify so that the bottom surface is hotter than the top.

So, with no greenhouse gasses in the atmosphere: $$F_{sun}=sigma T^4,$$ where $F_{sun}$ is the average, over the whole surface of the earth, of the heat flux absorbed from the sun (roughly $240 W/m^2$), $sigma$ is the Stefan-Boltzmann constant, $5.67 times 10^{-8}Wm^{-2}K^{-4}$, and T is the average temperature of the earth's surface. This results in $T=255K=-18^{circ}C$. Pretty chilly.

With a "perfect" greenhouse atmosphere, $$2F_{sun}=sigma T^4,$$ which results in $T=303K=30^{circ}C$, about 15 degrees warmer than the present average. You could reach a slightly higher average surface temperature--perhaps a couple of degrees--by assuming that the ice caps melt, lowering the albedo.

Looking at other planets in the solar system is a good way to get an indication of the role extreme variations of CO₂ has. Mercury is closest to the sun, and has a high maximum temperature (449 centigrade) but a very low minimum (-170 centigrade). It has practically no atmosphere, and virtually no CO₂. Venus, on the other hand - the planet between the Earth and Mercury - has a minimum temperature that is higher than Mercury's maximum (!) at 465 centigrade.... and also the same temperature for the maximum. It has 96% CO₂ in its thick atmosphere, keeping the planet very hot all of the time.

Temperatures from: http://earthguide.ucsd.edu/eoc/special_topics/teach/sp_climate_change/p_planet_temp.html